🐛 fix(app): 为稳定期首次连接增加瞬时网络重试保护 (#309 )

## 问题背景在 app 启动后等待 20s 以上，再手动触发数据库连接时，遇到瞬时网络错误（如 `no route to host`）会立即失败，用户体感为“没有重试”。相关讨论与上下文参考： - https://github.com/Syngnat/GoNavi/pull/294 ## 问题描述此前重试保护逻辑主要以“应用启动窗口（20s）”为边界： - 启动窗口内：瞬时网络失败会自动重试 - 启动窗口外：即使是瞬时网络失败也不重试这导致“用户首次手动连接发生在稳定期”时，行为与预期不一致。 ## 本地复现关键日志（节选） ```log 2026/03/27 15:21:04.792462 [INFO] 应用启动完成（首次连接保护窗口=20s，最多重试=4 次） 2026/03/27 15:22:29.196794 [INFO] 获取数据库连接：... 启动阶段=稳定期(age=1m24.405s) 2026/03/27 15:22:29.208920 [ERROR] 建立数据库连接失败：... connect: no route to host 2026/03/27 15:22:29.212453 [ERROR] DBGetDatabases 获取连接失败：... connect: no route to host 2026/03/27 16:07:45.463959 [INFO] 获取数据库连接：... 启动阶段=稳定期(age=21s) 2026/03/27 16:07:45.470744 [ERROR] 建立数据库连接失败：... connect: no route to host 2026/03/27 16:07:45.473604 [WARN] 检测到瞬时网络失败，准备重试连接：... 尝试=1/4 延迟=800ms 2026/03/27 16:07:46.277658 [INFO] 获取数据库连接：... 启动阶段=稳定期(age=21.814s) 2026/03/27 16:07:46.281761 [INFO] 创建数据库驱动实例：... 尝试=2/4 2026/03/27 16:07:46.284741 [ERROR] 建立数据库连接失败：... connect: no route to host 2026/03/27 16:20:59.298636 [INFO] 应用启动完成（首次连接保护窗口=20s，最多重试=4 次） 2026/03/27 16:23:26.180978 [INFO] 获取数据库连接：... 启动阶段=稳定期(age=2m26.883s) 2026/03/27 16:23:26.201478 [INFO] 数据库连接成功并写入缓存：... ``` ## 变更内容 - 调整连接重试判定逻辑： - 启动窗口内：保持原有重试预算（最多 4 次） - 启动窗口外：若为瞬时网络错误，补充一次保护重试（总计 2 次尝试） - 非瞬时错误（如认证失败）在稳定期不重试 - 日志文案泛化，避免“仅启动期”误导： - 数据库连接在重试后成功 - 检测到瞬时网络失败，准备重试连接 ## 测试与验证 ### 新增/更新单元测试覆盖以下场景： - 启动期瞬时错误重试并成功 - 稳定期瞬时错误重试一次并成功 - 稳定期瞬时错误持续失败时，仅重试一次后停止 - 稳定期非瞬时错误不重试 - 稳定期重试路径输出重试提示日志 - 启动期瞬时错误失败时使用完整重试预算 ### 本地执行： - go test ./internal/app -run StartupRetry -count=1 - go test ./internal/app -count=1 ### 影响范围 - 连接建立重试策略（internal/app/app.go） - 启动重试相关测试（internal/app/app_startup_connect_retry_test.go） ## 风险与回滚 - 风险：稳定期瞬时网络错误会增加一次重试等待（约 800ms） - 回滚：可回退本 PR 即恢复“仅启动窗口重试”的旧策略
🐛 fix(ai/provider/chat-ui): 修复千问 Coding Plan 预设与 Claude CLI 报错
2026-05-19 02:29:30 +08:00 · 2026-03-27 17:30:14 +08:00 · 2026-03-27 17:02:51 +08:00 · 2026-03-27 16:27:46 +08:00 · 2026-03-27 14:29:03 +08:00 · 2026-03-27 12:04:55 +08:00
31 changed files with 2044 additions and 158 deletions
--- a/frontend/src/components/AIChatPanel.tsx
+++ b/frontend/src/components/AIChatPanel.tsx
@@ -6,7 +6,6 @@ import { DBGetDatabases, DBGetTables } from '../../wailsjs/go/app/App';
 import type { OverlayWorkbenchTheme } from '../utils/overlayWorkbenchTheme';
 import { AIChatMessage, AIToolCall } from '../types';
 import { DownOutlined } from '@ant-design/icons';
-import { message as antdMessage } from 'antd';
 import './AIChatPanel.css';

 import { AIChatHeader } from './ai/AIChatHeader';
@@ -14,6 +13,12 @@ import { AIChatWelcome } from './ai/AIChatWelcome';
 import { AIMessageBubble } from './ai/AIMessageBubble';
 import { AIChatInput } from './ai/AIChatInput';
 import { AIHistoryDrawer } from './ai/AIHistoryDrawer';
+import type { AIComposerNotice } from '../utils/aiComposerNotice';
+import {
+    buildMissingModelNotice,
+    buildMissingProviderNotice,
+    buildModelFetchFailedNotice,
+} from '../utils/aiComposerNotice';

 interface AIChatPanelProps {
    width?: number;
@@ -211,6 +216,7 @@ export const AIChatPanel: React.FC<AIChatPanelProps> = ({
    const [dynamicModels, setDynamicModels] = useState<string[]>([]);
    const [showScrollBottom, setShowScrollBottom] = useState(false);
    const [loadingModels, setLoadingModels] = useState(false);
+    const [composerNotice, setComposerNotice] = useState<AIComposerNotice | null>(null);
    const [panelWidth, setPanelWidth] = useState(width);
    const [isResizing, setIsResizing] = useState(false);
    const [historyOpen, setHistoryOpen] = useState(false);
@@ -224,9 +230,6 @@ export const AIChatPanel: React.FC<AIChatPanelProps> = ({
    const panelRef = useRef<HTMLDivElement>(null); // 面板 DOM ref，用于拖拽时直接操作宽度
    const dragWidthRef = useRef(0); // 拖拽过程中的实时宽度（不触发 React 重渲染）

-    // 面板内部 toast 通知（不在屏幕顶部，而在面板容器内显示）
-    const [messageApi, messageContextHolder] = antdMessage.useMessage({ getContainer: () => panelRef.current || document.body });
-
    const aiChatHistory = useStore(state => state.aiChatHistory);
    const aiActiveSessionId = useStore(state => state.aiActiveSessionId);
    const createNewAISession = useStore(state => state.createNewAISession);
@@ -336,6 +339,7 @@ export const AIChatPanel: React.FC<AIChatPanelProps> = ({
    useEffect(() => {
        const handler = () => {
            setDynamicModels([]);
+            setComposerNotice(null);
            activeProviderIdRef.current = null;
            loadActiveProvider();
        };
@@ -350,6 +354,7 @@ export const AIChatPanel: React.FC<AIChatPanelProps> = ({
            const payload = { ...activeProvider, model: val };
            await Service?.AISaveProvider?.(payload);
            setActiveProvider(payload);
+            setComposerNotice(null);
        } catch (e) { console.warn('Failed to update provider model', e); }
    };

@@ -358,33 +363,45 @@ export const AIChatPanel: React.FC<AIChatPanelProps> = ({
    useEffect(() => {
        if (activeProvider?.id && activeProvider.id !== activeProviderIdRef.current) {
            setDynamicModels([]);
+            setComposerNotice(null);
            activeProviderIdRef.current = activeProvider.id;
        }
        // 供应商被删除后 activeProvider 变为 null，此时也必须清空残留模型
        if (!activeProvider) {
            setDynamicModels([]);
+            setComposerNotice(null);
            activeProviderIdRef.current = null;
        }
    }, [activeProvider?.id, activeProvider]);

+    useEffect(() => {
+        if (activeProvider?.model && String(activeProvider.model).trim()) {
+            setComposerNotice(null);
+        }
+    }, [activeProvider?.model]);
+

    // dynamicModels 仅在内存中使用，不再写回供应商配置，避免污染静态 models 列表

    const fetchDynamicModels = useCallback(async () => {
        try {
            setLoadingModels(true);
+            setComposerNotice(null);
            const Service = (window as any).go?.aiservice?.Service;
            if (!Service) return;
            const result = await Service.AIListModels?.();
            if (result?.success && Array.isArray(result.models) && result.models.length > 0) {
                const sortedModels = [...result.models].sort((a, b) => a.localeCompare(b));
                setDynamicModels(sortedModels);
+                setComposerNotice(null);
            } else if (result && !result.success) {
-                messageApi.warning(result.error || '获取模型列表失败，可手动输入模型名称');
+                setDynamicModels([]);
+                setComposerNotice(buildModelFetchFailedNotice(result.error));
            }
        } catch (e: any) {
            console.warn('Failed to fetch models', e);
-            messageApi.warning('获取模型列表失败: ' + (e?.message || '未知错误'));
+            setDynamicModels([]);
+            setComposerNotice(buildModelFetchFailedNotice('获取模型列表失败：' + (e?.message || '未知错误')));
        } finally {
            setLoadingModels(false);
        }
@@ -1030,13 +1047,14 @@ SELECT * FROM users WHERE status = 1;

        // 前置校验：必须配置供应商且选择模型后才能发送
        if (!activeProvider) {
-            messageApi.warning('请先在 AI 设置中配置供应商');
+            setComposerNotice(buildMissingProviderNotice());
            return;
        }
        if (!activeProvider.model || !activeProvider.model.trim()) {
-            messageApi.warning('请先选择模型 ID（点击工具栏的模型下拉框选择）');
+            setComposerNotice(buildMissingModelNotice());
            return;
        }
+        setComposerNotice(null);

        toolCallRoundRef.current = 0; // 重置工具调用轮次计数
        nudgeCountRef.current = 0;     // 重置催促计数
@@ -1258,7 +1276,6 @@ SELECT * FROM users WHERE status = 1;

    return (
        <div ref={panelRef} className="ai-chat-panel" style={{ width: panelWidth, background: bgColor || 'transparent', color: textColor, borderLeft: overlayTheme.shellBorder, position: 'relative' }}>
-            {messageContextHolder}
            <div className={`ai-resize-handle${isResizing ? ' active' : ''}`} onMouseDown={handleResizeStart} />
            
            {isResizing && panelRect.current && createPortal(
@@ -1366,6 +1383,7 @@ SELECT * FROM users WHERE status = 1;
                activeProvider={activeProvider}
                dynamicModels={dynamicModels}
                loadingModels={loadingModels}
+                composerNotice={composerNotice}
                onModelChange={handleModelChange}
                onFetchModels={fetchDynamicModels}
                textareaRef={textareaRef}
--- a/frontend/src/components/AISettingsModal.tsx
+++ b/frontend/src/components/AISettingsModal.tsx
@@ -2,6 +2,24 @@ import React, { useState, useEffect, useCallback, useRef } from 'react';
 import { Modal, Button, Input, Select, Form, message as antdMessage, Tooltip, Tabs, Space, Popconfirm, Slider } from 'antd';
 import { PlusOutlined, DeleteOutlined, EditOutlined, CheckOutlined, ApiOutlined, SafetyCertificateOutlined, RobotOutlined, ThunderboltOutlined, CloudOutlined, ExperimentOutlined, KeyOutlined, LinkOutlined, AppstoreOutlined, ToolOutlined } from '@ant-design/icons';
 import type { AIProviderConfig, AIProviderType, AISafetyLevel, AIContextLevel } from '../types';
+import {
+    getProviderFingerprint,
+    getProviderHostname,
+    matchQwenPresetKey,
+    QWEN_BAILIAN_ANTHROPIC_BASE_URL,
+    QWEN_CODING_PLAN_ANTHROPIC_BASE_URL,
+    QWEN_CODING_PLAN_MODELS,
+    resolvePresetBaseURL,
+    resolvePresetModelSelection,
+    resolvePresetTransport,
+} from '../utils/aiProviderPresets';
+import {
+    PROVIDER_PRESET_CARD_BASE_STYLE,
+    PROVIDER_PRESET_CARD_CONTENT_STYLE,
+    PROVIDER_PRESET_CARD_DESCRIPTION_STYLE,
+    PROVIDER_PRESET_GRID_STYLE,
+    PROVIDER_PRESET_CARD_TITLE_STYLE,
+} from '../utils/aiSettingsPresetLayout';

 import type { OverlayWorkbenchTheme } from '../utils/overlayWorkbenchTheme';

@@ -20,6 +38,7 @@ interface ProviderPreset {
    desc: string;
    color: string;
    backendType: AIProviderType;
+    fixedApiFormat?: string;
    defaultBaseUrl: string;
    defaultModel: string;
    models: string[];
@@ -28,12 +47,14 @@ interface ProviderPreset {
 const PROVIDER_PRESETS: ProviderPreset[] = [
    { key: 'openai', label: 'OpenAI', icon: <ApiOutlined />, desc: 'GPT-5.4 / 5.3 系列', color: '#10b981', backendType: 'openai', defaultBaseUrl: 'https://api.openai.com/v1', defaultModel: 'gpt-4o', models: [] },
    { key: 'deepseek', label: 'DeepSeek', icon: <ThunderboltOutlined />, desc: 'DeepSeek-V4 / R1', color: '#3b82f6', backendType: 'openai', defaultBaseUrl: 'https://api.deepseek.com/v1', defaultModel: 'deepseek-chat', models: [] },
-    { key: 'qwen', label: '通义千问', icon: <CloudOutlined />, desc: 'Qwen3.5 / Qwen3 系列', color: '#6366f1', backendType: 'openai', defaultBaseUrl: 'https://dashscope.aliyuncs.com/compatible-mode/v1', defaultModel: 'qwen-max', models: [] },
+    { key: 'qwen-bailian', label: '通义千问（百炼通用）', icon: <CloudOutlined />, desc: '百炼 Anthropic 兼容 / 模型从远端拉取', color: '#6366f1', backendType: 'anthropic', defaultBaseUrl: QWEN_BAILIAN_ANTHROPIC_BASE_URL, defaultModel: '', models: [] },
+    { key: 'qwen-coding-plan', label: '通义千问（Coding Plan）', icon: <CloudOutlined />, desc: 'Claude Code CLI 代理链路 / 使用官方支持模型清单', color: '#4f46e5', backendType: 'custom', fixedApiFormat: 'claude-cli', defaultBaseUrl: QWEN_CODING_PLAN_ANTHROPIC_BASE_URL, defaultModel: '', models: QWEN_CODING_PLAN_MODELS },
    { key: 'zhipu', label: '智谱 GLM', icon: <ExperimentOutlined />, desc: 'GLM-5 / GLM-5-Turbo', color: '#0ea5e9', backendType: 'openai', defaultBaseUrl: 'https://open.bigmodel.cn/api/paas/v4', defaultModel: 'glm-4', models: [] },
    { key: 'moonshot', label: 'Kimi', icon: <ExperimentOutlined />, desc: 'Kimi K2.5 (Anthropic 兼容)', color: '#0d9488', backendType: 'anthropic', defaultBaseUrl: 'https://api.moonshot.cn/anthropic', defaultModel: 'moonshot-v1-8k', models: [] },
    { key: 'anthropic', label: 'Claude', icon: <ExperimentOutlined />, desc: 'Claude Opus/Sonnet', color: '#d97706', backendType: 'anthropic', defaultBaseUrl: 'https://api.anthropic.com', defaultModel: 'claude-3-5-sonnet-20241022', models: [] },
    { key: 'gemini', label: 'Gemini', icon: <CloudOutlined />, desc: 'Gemini 3.1 / 2.5 系列', color: '#059669', backendType: 'gemini', defaultBaseUrl: 'https://generativelanguage.googleapis.com', defaultModel: 'gemini-2.5-flash', models: [] },
-    { key: 'volcengine', label: '火山引擎', icon: <CloudOutlined />, desc: '火山方舟 / 豆包大模型', color: '#0ea5e9', backendType: 'openai', defaultBaseUrl: 'https://ark.cn-beijing.volces.com/api/v3', defaultModel: 'ep-xxxxxx', models: [] },
+    { key: 'volcengine-ark', label: '火山方舟', icon: <CloudOutlined />, desc: 'Ark 通用推理 / 豆包模型', color: '#0ea5e9', backendType: 'openai', defaultBaseUrl: 'https://ark.cn-beijing.volces.com/api/v3', defaultModel: '', models: [] },
+    { key: 'volcengine-coding', label: '火山 Coding Plan', icon: <CloudOutlined />, desc: 'Ark Code / Coding Plan', color: '#0284c7', backendType: 'openai', defaultBaseUrl: 'https://ark.cn-beijing.volces.com/api/coding/v3', defaultModel: '', models: [] },
    { key: 'minimax', label: 'MiniMax', icon: <ExperimentOutlined />, desc: 'M2.7 / M2.5 系列 (Anthropic 兼容)', color: '#e11d48', backendType: 'anthropic', defaultBaseUrl: 'https://api.minimaxi.com/anthropic', defaultModel: 'MiniMax-M2.7', models: ['MiniMax-M2.7', 'MiniMax-M2.7-highspeed', 'MiniMax-M2.5', 'MiniMax-M2.5-highspeed', 'MiniMax-M2.1', 'MiniMax-M2.1-highspeed', 'MiniMax-M2'] },
    { key: 'ollama', label: 'Ollama', icon: <AppstoreOutlined />, desc: '本地部署开源模型', color: '#78716c', backendType: 'openai', defaultBaseUrl: 'http://localhost:11434/v1', defaultModel: 'llama3', models: [] },
    { key: 'custom', label: '自定义', icon: <AppstoreOutlined />, desc: '自定义 API 端点', color: '#64748b', backendType: 'custom', defaultBaseUrl: '', defaultModel: '', models: [] },
@@ -41,16 +62,21 @@ const PROVIDER_PRESETS: ProviderPreset[] = [

 const findPreset = (key: string): ProviderPreset => PROVIDER_PRESETS.find(p => p.key === key) || PROVIDER_PRESETS[PROVIDER_PRESETS.length - 1];

-const getProviderHostname = (raw?: string): string => {
-    if (!raw) return '';
-    try {
-        return new URL(raw).hostname.toLowerCase();
-    } catch {
-        return '';
-    }
-};
-
 const matchProviderPreset = (provider: Pick<AIProviderConfig, 'type' | 'baseUrl'>): ProviderPreset => {
+    const qwenPresetKey = matchQwenPresetKey(provider);
+    if (qwenPresetKey) {
+        return findPreset(qwenPresetKey);
+    }
+    const fingerprint = getProviderFingerprint(provider.baseUrl);
+    const exactPreset = PROVIDER_PRESETS.find(pr =>
+        pr.backendType === provider.type
+        && fingerprint !== ''
+        && fingerprint === getProviderFingerprint(pr.defaultBaseUrl)
+    );
+    if (exactPreset) {
+        return exactPreset;
+    }
+
    const host = getProviderHostname(provider.baseUrl);
    if (host.endsWith('moonshot.cn')) {
        return findPreset('moonshot');
@@ -143,11 +169,22 @@ const AISettingsModal: React.FC<AISettingsModalProps> = ({ open, onClose, darkMo
    const handleEditProvider = (p: AIProviderConfig) => {
        // 尝试根据 baseUrl 和 type 推断 preset
        const matchedPreset = matchProviderPreset(p);
+        const resolvedTransport = resolvePresetTransport({
+            presetBackendType: matchedPreset.backendType,
+            presetFixedApiFormat: matchedPreset.fixedApiFormat,
+            valuesApiFormat: p.apiFormat,
+        });
        setEditingProvider(p);
        setIsEditing(true);
        setTestStatus('idle');
        form.resetFields();
-        form.setFieldsValue({ ...p, type: matchedPreset.backendType, models: p.models || [], presetKey: matchedPreset.key, apiFormat: p.apiFormat || 'openai' });
+        form.setFieldsValue({
+            ...p,
+            type: resolvedTransport.type,
+            models: p.models || [],
+            presetKey: matchedPreset.key,
+            apiFormat: resolvedTransport.apiFormat || p.apiFormat || 'openai',
+        });
    };

    const handleDeleteProvider = async (id: string) => {
@@ -179,24 +216,38 @@ const AISettingsModal: React.FC<AISettingsModalProps> = ({ open, onClose, darkMo
            const Service = (window as any).go?.aiservice?.Service;
            
            // 构建 payload，处理 model/models 逻辑
-            const isCustomLike = values.presetKey === 'custom' || values.presetKey === 'ollama';
            const preset = findPreset(values.presetKey);
-            const resolvedModels = isCustomLike ? (values.models || []) : preset.models;
-            const fallbackModel = resolvedModels.length > 0 ? resolvedModels[0] : '';
-            const finalModel = isCustomLike ? fallbackModel : (values.model || fallbackModel);
+            const isCustomLike = values.presetKey === 'custom' || values.presetKey === 'ollama';
+            const { model: finalModel, models: resolvedModels } = resolvePresetModelSelection({
+                presetKey: values.presetKey,
+                presetDefaultModel: preset.defaultModel,
+                presetModels: preset.models,
+                valuesModel: values.model,
+                customModels: values.models,
+            });
            // 内置供应商自动使用 preset label 作为名称
            const finalName = isCustomLike ? (values.name || preset.label) : preset.label;
            
-            const finalBaseUrl = values.baseUrl || preset.defaultBaseUrl;
+            const finalBaseUrl = resolvePresetBaseURL({
+                presetKey: values.presetKey,
+                presetDefaultBaseUrl: preset.defaultBaseUrl,
+                valuesBaseUrl: values.baseUrl,
+            });
+            const resolvedTransport = resolvePresetTransport({
+                presetBackendType: preset.backendType,
+                presetFixedApiFormat: preset.fixedApiFormat,
+                valuesApiFormat: values.apiFormat,
+            });
            
            const payload = { 
                ...editingProvider, 
                ...values, 
+                ...resolvedTransport,
                name: finalName,
                model: finalModel,
                models: resolvedModels,
                baseUrl: finalBaseUrl,
-                apiFormat: values.apiFormat || 'openai',
+                apiFormat: resolvedTransport.apiFormat,
            };
            // 后端 AISaveProvider 统一处理新增和更新，返回 void，失败抛异常
            await Service?.AISaveProvider?.(payload);
@@ -240,8 +291,34 @@ const AISettingsModal: React.FC<AISettingsModalProps> = ({ open, onClose, darkMo
            setTestStatus('idle');
            const Service = (window as any).go?.aiservice?.Service;
            const preset = findPreset(values.presetKey || 'openai');
-            const finalBaseUrl = values.baseUrl || preset.defaultBaseUrl;
-            const res = await Service?.AITestProvider?.({ ...values, baseUrl: finalBaseUrl, maxTokens: Number(values.maxTokens) || 4096, temperature: Number(values.temperature) ?? 0.7 });
+            const finalBaseUrl = resolvePresetBaseURL({
+                presetKey: values.presetKey || 'openai',
+                presetDefaultBaseUrl: preset.defaultBaseUrl,
+                valuesBaseUrl: values.baseUrl,
+            });
+            const { model: finalModel, models: resolvedModels } = resolvePresetModelSelection({
+                presetKey: values.presetKey || 'openai',
+                presetDefaultModel: preset.defaultModel,
+                presetModels: preset.models,
+                valuesModel: values.model,
+                customModels: values.models,
+            });
+            const resolvedTransport = resolvePresetTransport({
+                presetBackendType: preset.backendType,
+                presetFixedApiFormat: preset.fixedApiFormat,
+                valuesApiFormat: values.apiFormat,
+            });
+            const res = await Service?.AITestProvider?.({
+                ...editingProvider,
+                ...values,
+                ...resolvedTransport,
+                baseUrl: finalBaseUrl,
+                model: finalModel,
+                models: resolvedModels,
+                maxTokens: Number(values.maxTokens) || 4096,
+                temperature: Number(values.temperature) ?? 0.7,
+                apiFormat: resolvedTransport.apiFormat,
+            });
            if (res?.success) { setTestStatus('success'); void messageApi.success('连接成功'); }
            else { setTestStatus('error'); void messageApi.error(`测试失败: ${res?.message || '未知错误'}`); }
        } catch (e: any) { setTestStatus('error'); void messageApi.error(e?.message || '测试失败'); }
@@ -250,9 +327,15 @@ const AISettingsModal: React.FC<AISettingsModalProps> = ({ open, onClose, darkMo

    const handlePresetChange = (presetKey: string) => {
        const preset = findPreset(presetKey);
+        const resolvedTransport = resolvePresetTransport({
+            presetBackendType: preset.backendType,
+            presetFixedApiFormat: preset.fixedApiFormat,
+            valuesApiFormat: form.getFieldValue('apiFormat'),
+        });
        form.setFieldsValue({
            presetKey,
-            type: preset.backendType,
+            type: resolvedTransport.type,
+            apiFormat: resolvedTransport.apiFormat || 'openai',
            baseUrl: preset.defaultBaseUrl,
            model: preset.defaultModel,
        });
@@ -307,7 +390,7 @@ const AISettingsModal: React.FC<AISettingsModalProps> = ({ open, onClose, darkMo
                            <div style={{ fontSize: 12, color: overlayTheme.mutedText, marginTop: 4, display: 'flex', alignItems: 'center', gap: 6 }}>
                                <span>{matchedPreset.label}</span>
                                <span style={{ opacity: 0.4 }}>·</span>
-                                <span style={{ fontFamily: 'monospace', fontSize: 12 }}>{p.model}</span>
+                                <span style={{ fontFamily: 'monospace', fontSize: 12 }}>{p.model || '未选择模型'}</span>
                            </div>
                        </div>
                        <Space size={2}>
@@ -353,25 +436,24 @@ const AISettingsModal: React.FC<AISettingsModalProps> = ({ open, onClose, darkMo
                            <AppstoreOutlined style={{ fontSize: 14 }} /> 服务类型
                        </div>
                        <Form.Item name="presetKey" noStyle>
-                            <div style={{ display: 'grid', gridTemplateColumns: '1fr 1fr 1fr', gap: 6 }}>
+                            <div style={PROVIDER_PRESET_GRID_STYLE}>
                                {PROVIDER_PRESETS.map(pt => (
                                    <div key={pt.key} onClick={() => { form.setFieldValue('presetKey', pt.key); handlePresetChange(pt.key); }}
                                        style={{
-                                            padding: '12px 14px', borderRadius: 12, cursor: 'pointer', transition: 'all 0.2s ease',
+                                            ...PROVIDER_PRESET_CARD_BASE_STYLE,
                                            border: `1.5px solid ${presetKeyFromForm === pt.key ? overlayTheme.selectedText : 'transparent'}`,
                                            background: presetKeyFromForm === pt.key ? overlayTheme.selectedBg : (darkMode ? 'rgba(255,255,255,0.02)' : 'rgba(255,255,255,0.72)'),
                                            boxShadow: presetKeyFromForm === pt.key ? 'none' : (darkMode ? 'inset 0 0 0 1px rgba(255,255,255,0.028)' : 'inset 0 0 0 1px rgba(16,24,40,0.03)'),
-                                            display: 'flex', alignItems: 'flex-start', gap: 10,
                                        }}>
                                        <div style={{
                                            color: presetKeyFromForm === pt.key ? overlayTheme.iconColor : overlayTheme.mutedText,
-                                            fontSize: 18, marginTop: 2, transition: 'all 0.2s ease',
+                                            fontSize: 18, marginTop: 2, transition: 'all 0.2s ease', flexShrink: 0,
                                        }}>
                                            {pt.icon}
                                        </div>
-                                        <div>
-                                            <div style={{ fontSize: 13, fontWeight: 700, color: overlayTheme.titleText, lineHeight: 1.3 }}>{pt.label}</div>
-                                            <div style={{ fontSize: 12, color: overlayTheme.mutedText, marginTop: 4, lineHeight: 1.4 }}>{pt.desc}</div>
+                                        <div style={PROVIDER_PRESET_CARD_CONTENT_STYLE}>
+                                            <div style={{ ...PROVIDER_PRESET_CARD_TITLE_STYLE, fontSize: 13, fontWeight: 700, color: overlayTheme.titleText, lineHeight: 1.3 }}>{pt.label}</div>
+                                            <div style={{ ...PROVIDER_PRESET_CARD_DESCRIPTION_STYLE, fontSize: 12, color: overlayTheme.mutedText, lineHeight: 1.4 }}>{pt.desc}</div>
                                        </div>
                                    </div>
                                ))}
--- a/frontend/src/components/ai/AIChatInput.notice.test.tsx
+++ b/frontend/src/components/ai/AIChatInput.notice.test.tsx
@@ -0,0 +1,61 @@
+import React from 'react';
+import { renderToStaticMarkup } from 'react-dom/server';
+import { describe, expect, it, vi } from 'vitest';
+
+import { AIChatInput } from './AIChatInput';
+import { buildOverlayWorkbenchTheme } from '../../utils/overlayWorkbenchTheme';
+
+vi.mock('../../store', () => ({
+  useStore: (selector: (state: any) => any) => selector({
+    aiContexts: {},
+    addAIContext: vi.fn(),
+    removeAIContext: vi.fn(),
+  }),
+}));
+
+vi.mock('../../../wailsjs/go/app/App', () => ({
+  DBGetTables: vi.fn(),
+  DBShowCreateTable: vi.fn(),
+  DBGetDatabases: vi.fn(),
+}));
+
+describe('AIChatInput notice layout', () => {
+  it('renders the composer notice above the input editor', () => {
+    const markup = renderToStaticMarkup(
+      <AIChatInput
+        input=""
+        setInput={() => {}}
+        draftImages={[]}
+        setDraftImages={() => {}}
+        sending={false}
+        onSend={() => {}}
+        onStop={() => {}}
+        handleKeyDown={() => {}}
+        activeConnName=""
+        activeContext={null}
+        activeProvider={{ model: '', models: [] }}
+        dynamicModels={[]}
+        loadingModels={false}
+        composerNotice={{
+          tone: 'error',
+          title: '模型列表加载失败',
+          description: '请检查供应商入口和 API Key。',
+        }}
+        onModelChange={() => {}}
+        onFetchModels={() => {}}
+        textareaRef={React.createRef<HTMLTextAreaElement>()}
+        darkMode={false}
+        textColor="#162033"
+        mutedColor="rgba(16,24,40,0.55)"
+        overlayTheme={buildOverlayWorkbenchTheme(false)}
+      />
+    );
+
+    const noticeIndex = markup.indexOf('data-ai-chat-composer-notice="true"');
+    const inputIndex = markup.indexOf('data-ai-chat-composer-input="true"');
+
+    expect(noticeIndex).toBeGreaterThanOrEqual(0);
+    expect(inputIndex).toBeGreaterThanOrEqual(0);
+    expect(noticeIndex).toBeLessThan(inputIndex);
+  });
+});
--- a/frontend/src/components/ai/AIChatInput.tsx
+++ b/frontend/src/components/ai/AIChatInput.tsx
@@ -1,9 +1,10 @@
 import React from 'react';
 import { Input, Select, AutoComplete, Tooltip, Modal, Checkbox, Spin, message, Button, Tag } from 'antd';
-import { DatabaseOutlined, SendOutlined, TableOutlined, SearchOutlined, PictureOutlined } from '@ant-design/icons';
+import { DatabaseOutlined, SendOutlined, TableOutlined, SearchOutlined, PictureOutlined, ExclamationCircleFilled } from '@ant-design/icons';
 import { useStore } from '../../store';
 import { DBGetTables, DBShowCreateTable, DBGetDatabases } from '../../../wailsjs/go/app/App';
 import type { OverlayWorkbenchTheme } from '../../utils/overlayWorkbenchTheme';
+import type { AIComposerNotice } from '../../utils/aiComposerNotice';

 interface AIChatInputProps {
    input: string;
@@ -19,6 +20,7 @@ interface AIChatInputProps {
    activeProvider: any;
    dynamicModels: string[];
    loadingModels: boolean;
+    composerNotice?: AIComposerNotice | null;
    onModelChange: (val: string) => void;
    onFetchModels: () => void;
    textareaRef: React.RefObject<HTMLTextAreaElement>;
@@ -33,6 +35,7 @@ interface AIChatInputProps {
 export const AIChatInput: React.FC<AIChatInputProps> = ({
    input, setInput, draftImages, setDraftImages, sending, onSend, onStop, handleKeyDown,
    activeConnName, activeContext, activeProvider, dynamicModels, loadingModels,
+    composerNotice,
    onModelChange, onFetchModels, textareaRef, darkMode, textColor, mutedColor, overlayTheme,
    contextUsageChars, maxContextChars
 }) => {
@@ -67,6 +70,33 @@ export const AIChatInput: React.FC<AIChatInputProps> = ({

    const filteredTables = contextTables.filter(t => t.name.toLowerCase().includes(searchText.toLowerCase()));
    const [contextExpanded, setContextExpanded] = React.useState(false);
+    const composerNoticePalette = React.useMemo(() => {
+        if (composerNotice?.tone === 'error') {
+            return darkMode
+                ? {
+                    background: 'rgba(255,120,117,0.12)',
+                    borderColor: 'rgba(255,120,117,0.24)',
+                    iconColor: '#ff7875',
+                }
+                : {
+                    background: 'rgba(255,77,79,0.08)',
+                    borderColor: 'rgba(255,77,79,0.16)',
+                    iconColor: '#ff4d4f',
+                };
+        }
+
+        return darkMode
+            ? {
+                background: 'rgba(250,173,20,0.12)',
+                borderColor: 'rgba(250,173,20,0.22)',
+                iconColor: '#ffd666',
+            }
+            : {
+                background: 'rgba(250,173,20,0.08)',
+                borderColor: 'rgba(250,173,20,0.18)',
+                iconColor: '#d48806',
+            };
+    }, [composerNotice, darkMode]);

    // Slash commands
    const [showSlashMenu, setShowSlashMenu] = React.useState(false);
@@ -258,7 +288,31 @@ export const AIChatInput: React.FC<AIChatInputProps> = ({
                        </div>
                    ))}
                </div>
-                <div style={{ position: 'relative' }}>
+                {composerNotice && (
+                    <div
+                        data-ai-chat-composer-notice="true"
+                        style={{
+                            display: 'flex',
+                            alignItems: 'flex-start',
+                            gap: 8,
+                            padding: '8px 10px',
+                            borderRadius: 12,
+                            background: composerNoticePalette.background,
+                            border: `1px solid ${composerNoticePalette.borderColor}`,
+                        }}
+                    >
+                        <ExclamationCircleFilled style={{ color: composerNoticePalette.iconColor, fontSize: 14, marginTop: 1, flexShrink: 0 }} />
+                        <div style={{ minWidth: 0 }}>
+                            <div style={{ fontSize: 12, fontWeight: 600, color: textColor, lineHeight: 1.4 }}>
+                                {composerNotice.title}
+                            </div>
+                            <div style={{ fontSize: 11, color: mutedColor, lineHeight: 1.5, marginTop: 2, wordBreak: 'break-word' }}>
+                                {composerNotice.description}
+                            </div>
+                        </div>
+                    </div>
+                )}
+                <div data-ai-chat-composer-input="true" style={{ position: 'relative' }}>
                    {showSlashMenu && filteredSlashCmds.length > 0 && (
                        <div style={{
                            position: 'absolute', bottom: '100%', left: 0, right: 0, marginBottom: 4,
@@ -354,9 +408,13 @@ export const AIChatInput: React.FC<AIChatInputProps> = ({
                            <Select
                                size="small"
                                variant="filled"
-                                value={activeProvider.model || (dynamicModels.length > 0 ? dynamicModels[0] : activeProvider.models?.[0])}
+                                value={activeProvider.model || undefined}
                                onChange={onModelChange}
-                                onDropdownVisibleChange={(open) => { if (open && dynamicModels.length === 0) onFetchModels(); }}
+                                onDropdownVisibleChange={(open) => {
+                                    if (open && dynamicModels.length === 0 && (activeProvider.models || []).length === 0) {
+                                        onFetchModels();
+                                    }
+                                }}
                                loading={loadingModels}
                                options={(dynamicModels.length > 0 ? dynamicModels : (activeProvider.models || [])).map((m: string) => ({ label: m, value: m }))}
                                style={{ width: 130, fontSize: 11, background: 'transparent' }}
--- a/frontend/src/types.ts
+++ b/frontend/src/types.ts
@@ -204,7 +204,7 @@ export interface AIProviderConfig {
  baseUrl: string;
  model: string;
  models?: string[];
-  apiFormat?: string; // custom 专用: openai | anthropic | gemini
+  apiFormat?: string; // custom 专用: openai | anthropic | gemini | claude-cli
  headers?: Record<string, string>;
  maxTokens: number;
  temperature: number;
@@ -243,4 +243,3 @@ export interface AISafetyResult {
  requiresConfirm: boolean;
  warningMessage?: string;
 }
-
--- a/frontend/src/utils/aiComposerNotice.test.ts
+++ b/frontend/src/utils/aiComposerNotice.test.ts
@@ -0,0 +1,33 @@
+import { describe, expect, it } from 'vitest';
+
+import {
+  buildModelFetchFailedNotice,
+  buildMissingModelNotice,
+  buildMissingProviderNotice,
+} from './aiComposerNotice';
+
+describe('ai composer notice helpers', () => {
+  it('builds a compact notice for missing provider', () => {
+    expect(buildMissingProviderNotice()).toEqual({
+      tone: 'warning',
+      title: '还没有可用供应商',
+      description: '先在 AI 设置里添加并启用一个模型供应商。',
+    });
+  });
+
+  it('builds a compact notice for missing model selection', () => {
+    expect(buildMissingModelNotice()).toEqual({
+      tone: 'warning',
+      title: '先选择一个模型',
+      description: '打开下方模型下拉并选择模型；如果列表为空，请检查供应商入口和 API Key。',
+    });
+  });
+
+  it('builds a readable inline notice for model fetch failures', () => {
+    expect(buildModelFetchFailedNotice('当前接口未返回可用模型')).toEqual({
+      tone: 'error',
+      title: '模型列表加载失败',
+      description: '当前接口未返回可用模型',
+    });
+  });
+});
--- a/frontend/src/utils/aiComposerNotice.ts
+++ b/frontend/src/utils/aiComposerNotice.ts
@@ -0,0 +1,27 @@
+export type AIComposerNoticeTone = 'warning' | 'error';
+
+export interface AIComposerNotice {
+  tone: AIComposerNoticeTone;
+  title: string;
+  description: string;
+}
+
+const defaultModelFetchFailedDescription = '请检查供应商入口、API Key 或账号权限，然后重新打开模型下拉。';
+
+export const buildMissingProviderNotice = (): AIComposerNotice => ({
+  tone: 'warning',
+  title: '还没有可用供应商',
+  description: '先在 AI 设置里添加并启用一个模型供应商。',
+});
+
+export const buildMissingModelNotice = (): AIComposerNotice => ({
+  tone: 'warning',
+  title: '先选择一个模型',
+  description: '打开下方模型下拉并选择模型；如果列表为空，请检查供应商入口和 API Key。',
+});
+
+export const buildModelFetchFailedNotice = (error?: string): AIComposerNotice => ({
+  tone: 'error',
+  title: '模型列表加载失败',
+  description: String(error || '').trim() || defaultModelFetchFailedDescription,
+});
--- a/frontend/src/utils/aiProviderPresets.test.ts
+++ b/frontend/src/utils/aiProviderPresets.test.ts
@@ -0,0 +1,111 @@
+import { describe, expect, it } from 'vitest';
+
+import {
+  matchQwenPresetKey,
+  QWEN_BAILIAN_MODELS_BASE_URL,
+  QWEN_CODING_PLAN_ANTHROPIC_BASE_URL,
+  QWEN_CODING_PLAN_MODELS,
+  resolvePresetBaseURL,
+  resolvePresetModelSelection,
+  resolvePresetTransport,
+} from './aiProviderPresets';
+
+describe('ai provider preset helpers', () => {
+  it('maps legacy Bailian compatible-mode URL back to the Bailian preset', () => {
+    expect(matchQwenPresetKey({
+      type: 'openai',
+      baseUrl: QWEN_BAILIAN_MODELS_BASE_URL,
+    })).toBe('qwen-bailian');
+  });
+
+  it('maps Coding Plan anthropic URL to the dedicated Coding Plan preset', () => {
+    expect(matchQwenPresetKey({
+      type: 'anthropic',
+      baseUrl: QWEN_CODING_PLAN_ANTHROPIC_BASE_URL,
+    })).toBe('qwen-coding-plan');
+  });
+
+  it('maps Coding Plan Claude CLI config back to the dedicated Coding Plan preset', () => {
+    expect(matchQwenPresetKey({
+      type: 'custom',
+      apiFormat: 'claude-cli',
+      baseUrl: QWEN_CODING_PLAN_ANTHROPIC_BASE_URL,
+    })).toBe('qwen-coding-plan');
+  });
+
+  it('does not keep a baked-in model list for the Coding Plan preset', () => {
+    expect(QWEN_CODING_PLAN_MODELS).toEqual([
+      'qwen3.5-plus',
+      'kimi-k2.5',
+      'glm-5',
+      'MiniMax-M2.5',
+      'qwen3-max-2026-01-23',
+      'qwen3-coder-next',
+      'qwen3-coder-plus',
+      'glm-4.7',
+    ]);
+  });
+
+  it('keeps built-in preset model empty when the preset intentionally requires an explicit selection', () => {
+    expect(resolvePresetModelSelection({
+      presetKey: 'qwen-coding-plan',
+      presetDefaultModel: '',
+      presetModels: QWEN_CODING_PLAN_MODELS,
+      valuesModel: '',
+      customModels: [],
+    })).toEqual({
+      model: '',
+      models: QWEN_CODING_PLAN_MODELS,
+    });
+  });
+
+  it('still falls back to the first configured model for custom-like presets', () => {
+    expect(resolvePresetModelSelection({
+      presetKey: 'custom',
+      presetDefaultModel: '',
+      presetModels: [],
+      valuesModel: '',
+      customModels: ['foo-model', 'bar-model'],
+    })).toEqual({
+      model: 'foo-model',
+      models: ['foo-model', 'bar-model'],
+    });
+  });
+
+  it('forces built-in presets back to their standard base URL when saving or testing', () => {
+    expect(resolvePresetBaseURL({
+      presetKey: 'qwen-bailian',
+      presetDefaultBaseUrl: 'https://dashscope.aliyuncs.com/apps/anthropic',
+      valuesBaseUrl: 'https://dashscope.aliyuncs.com/compatible-mode/v1',
+    })).toBe('https://dashscope.aliyuncs.com/apps/anthropic');
+  });
+
+  it('keeps the user-entered base URL for custom-like presets', () => {
+    expect(resolvePresetBaseURL({
+      presetKey: 'custom',
+      presetDefaultBaseUrl: '',
+      valuesBaseUrl: 'https://example-proxy.internal/v1',
+    })).toBe('https://example-proxy.internal/v1');
+  });
+
+  it('forces qwen coding plan to save as custom plus claude-cli', () => {
+    expect(resolvePresetTransport({
+      presetBackendType: 'custom',
+      presetFixedApiFormat: 'claude-cli',
+      valuesApiFormat: 'anthropic',
+    })).toEqual({
+      type: 'custom',
+      apiFormat: 'claude-cli',
+    });
+  });
+
+  it('keeps custom preset transport editable', () => {
+    expect(resolvePresetTransport({
+      presetBackendType: 'custom',
+      valuesApiFormat: 'gemini',
+    })).toEqual({
+      type: 'custom',
+      apiFormat: 'gemini',
+    });
+  });
+});
--- a/frontend/src/utils/aiProviderPresets.ts
+++ b/frontend/src/utils/aiProviderPresets.ts
@@ -0,0 +1,143 @@
+import type { AIProviderConfig, AIProviderType } from '../types';
+
+export const LEGACY_QWEN_BAILIAN_OPENAI_BASE_URL = 'https://dashscope.aliyuncs.com/compatible-mode/v1';
+export const LEGACY_QWEN_CODING_PLAN_OPENAI_BASE_URL = 'https://coding.dashscope.aliyuncs.com/v1';
+export const QWEN_BAILIAN_ANTHROPIC_BASE_URL = 'https://dashscope.aliyuncs.com/apps/anthropic';
+export const QWEN_CODING_PLAN_ANTHROPIC_BASE_URL = 'https://coding.dashscope.aliyuncs.com/apps/anthropic';
+export const QWEN_BAILIAN_MODELS_BASE_URL = LEGACY_QWEN_BAILIAN_OPENAI_BASE_URL;
+
+export const QWEN_CODING_PLAN_MODELS = [
+  'qwen3.5-plus',
+  'kimi-k2.5',
+  'glm-5',
+  'MiniMax-M2.5',
+  'qwen3-max-2026-01-23',
+  'qwen3-coder-next',
+  'qwen3-coder-plus',
+  'glm-4.7',
+];
+
+const CUSTOM_LIKE_PRESET_KEYS = new Set(['custom', 'ollama']);
+
+export interface ResolvePresetModelSelectionInput {
+  presetKey: string;
+  presetDefaultModel: string;
+  presetModels: string[];
+  valuesModel?: string;
+  customModels?: string[];
+}
+
+export interface ResolvePresetModelSelectionResult {
+  model: string;
+  models: string[];
+}
+
+export interface ResolvePresetBaseURLInput {
+  presetKey: string;
+  presetDefaultBaseUrl: string;
+  valuesBaseUrl?: string;
+}
+
+export interface ResolvePresetTransportInput {
+  presetBackendType: AIProviderType;
+  presetFixedApiFormat?: string;
+  valuesApiFormat?: string;
+}
+
+export interface ResolvePresetTransportResult {
+  type: AIProviderType;
+  apiFormat?: string;
+}
+
+export const getProviderHostname = (raw?: string): string => {
+  if (!raw) return '';
+  try {
+    return new URL(raw).hostname.toLowerCase();
+  } catch {
+    return '';
+  }
+};
+
+export const getProviderFingerprint = (raw?: string): string => {
+  if (!raw) return '';
+  try {
+    const url = new URL(raw);
+    const normalizedPath = url.pathname.replace(/\/+$/, '').toLowerCase();
+    return `${url.hostname.toLowerCase()}${normalizedPath}`;
+  } catch {
+    return '';
+  }
+};
+
+export const matchQwenPresetKey = (provider: Pick<AIProviderConfig, 'type' | 'baseUrl' | 'apiFormat'>): string | null => {
+  const fingerprint = getProviderFingerprint(provider.baseUrl);
+  const bailianFingerprints = new Set([
+    getProviderFingerprint(LEGACY_QWEN_BAILIAN_OPENAI_BASE_URL),
+    getProviderFingerprint(QWEN_BAILIAN_ANTHROPIC_BASE_URL),
+  ]);
+  if (fingerprint !== '' && bailianFingerprints.has(fingerprint)) {
+    return 'qwen-bailian';
+  }
+
+  const codingPlanFingerprints = new Set([
+    getProviderFingerprint(LEGACY_QWEN_CODING_PLAN_OPENAI_BASE_URL),
+    getProviderFingerprint(QWEN_CODING_PLAN_ANTHROPIC_BASE_URL),
+  ]);
+  if (fingerprint !== '' && codingPlanFingerprints.has(fingerprint)) {
+    return 'qwen-coding-plan';
+  }
+
+  return null;
+};
+
+export const resolvePresetModelSelection = ({
+  presetKey,
+  presetDefaultModel,
+  presetModels,
+  valuesModel,
+  customModels,
+}: ResolvePresetModelSelectionInput): ResolvePresetModelSelectionResult => {
+  const isCustomLike = CUSTOM_LIKE_PRESET_KEYS.has(presetKey);
+  const resolvedModels = isCustomLike ? (customModels || []) : presetModels;
+  const fallbackModel = resolvedModels.length > 0 ? resolvedModels[0] : '';
+  return {
+    models: resolvedModels,
+    model: isCustomLike ? (valuesModel || fallbackModel) : (valuesModel || presetDefaultModel),
+  };
+};
+
+export const resolvePresetBaseURL = ({
+  presetKey,
+  presetDefaultBaseUrl,
+  valuesBaseUrl,
+}: ResolvePresetBaseURLInput): string => {
+  if (CUSTOM_LIKE_PRESET_KEYS.has(presetKey)) {
+    return valuesBaseUrl || presetDefaultBaseUrl;
+  }
+  return presetDefaultBaseUrl;
+};
+
+export const resolvePresetTransport = ({
+  presetBackendType,
+  presetFixedApiFormat,
+  valuesApiFormat,
+}: ResolvePresetTransportInput): ResolvePresetTransportResult => {
+  if (presetFixedApiFormat) {
+    return {
+      type: presetBackendType,
+      apiFormat: presetFixedApiFormat,
+    };
+  }
+
+  if (presetBackendType === 'custom') {
+    return {
+      type: presetBackendType,
+      apiFormat: valuesApiFormat || 'openai',
+    };
+  }
+
+  return {
+    type: presetBackendType,
+    apiFormat: undefined,
+  };
+};
--- a/frontend/src/utils/aiSettingsPresetLayout.test.ts
+++ b/frontend/src/utils/aiSettingsPresetLayout.test.ts
@@ -0,0 +1,56 @@
+import { describe, expect, it } from 'vitest';
+
+import {
+  PROVIDER_PRESET_CARD_BASE_STYLE,
+  PROVIDER_PRESET_CARD_CONTENT_STYLE,
+  PROVIDER_PRESET_CARD_DESCRIPTION_STYLE,
+  PROVIDER_PRESET_GRID_STYLE,
+  PROVIDER_PRESET_CARD_TITLE_STYLE,
+} from './aiSettingsPresetLayout';
+
+describe('ai settings preset layout', () => {
+  it('uses a fixed grid auto row height so provider bubbles stay visually consistent across rows', () => {
+    expect(PROVIDER_PRESET_GRID_STYLE).toMatchObject({
+      display: 'grid',
+      gridTemplateColumns: 'repeat(3, minmax(0, 1fr))',
+      gap: 6,
+      gridAutoRows: '96px',
+      alignItems: 'stretch',
+    });
+  });
+
+  it('stretches each provider card to fill the row height', () => {
+    expect(PROVIDER_PRESET_CARD_BASE_STYLE).toMatchObject({
+      display: 'flex',
+      alignItems: 'flex-start',
+      gap: 10,
+      height: '100%',
+      minHeight: '96px',
+      overflow: 'hidden',
+    });
+  });
+
+  it('keeps the text column compact instead of pinning the description to the bottom', () => {
+    expect(PROVIDER_PRESET_CARD_CONTENT_STYLE).toMatchObject({
+      minWidth: 0,
+      flex: 1,
+      display: 'flex',
+      flexDirection: 'column',
+    });
+
+    expect(PROVIDER_PRESET_CARD_DESCRIPTION_STYLE).toMatchObject({
+      marginTop: 4,
+      display: '-webkit-box',
+      WebkitLineClamp: 2,
+      WebkitBoxOrient: 'vertical',
+      overflow: 'hidden',
+    });
+
+    expect(PROVIDER_PRESET_CARD_TITLE_STYLE).toMatchObject({
+      display: '-webkit-box',
+      WebkitLineClamp: 2,
+      WebkitBoxOrient: 'vertical',
+      overflow: 'hidden',
+    });
+  });
+});
--- a/frontend/src/utils/aiSettingsPresetLayout.ts
+++ b/frontend/src/utils/aiSettingsPresetLayout.ts
@@ -0,0 +1,47 @@
+import type { CSSProperties } from 'react';
+
+export const PROVIDER_PRESET_CARD_HEIGHT = 96;
+
+export const PROVIDER_PRESET_GRID_STYLE: CSSProperties = {
+  display: 'grid',
+  gridTemplateColumns: 'repeat(3, minmax(0, 1fr))',
+  gap: 6,
+  gridAutoRows: `${PROVIDER_PRESET_CARD_HEIGHT}px`,
+  alignItems: 'stretch',
+};
+
+export const PROVIDER_PRESET_CARD_BASE_STYLE: CSSProperties = {
+  padding: '12px 14px',
+  borderRadius: 12,
+  cursor: 'pointer',
+  transition: 'all 0.2s ease',
+  display: 'flex',
+  alignItems: 'flex-start',
+  gap: 10,
+  height: '100%',
+  minHeight: `${PROVIDER_PRESET_CARD_HEIGHT}px`,
+  boxSizing: 'border-box',
+  overflow: 'hidden',
+};
+
+export const PROVIDER_PRESET_CARD_CONTENT_STYLE: CSSProperties = {
+  minWidth: 0,
+  flex: 1,
+  display: 'flex',
+  flexDirection: 'column',
+};
+
+export const PROVIDER_PRESET_CARD_DESCRIPTION_STYLE: CSSProperties = {
+  marginTop: 4,
+  display: '-webkit-box',
+  WebkitLineClamp: 2,
+  WebkitBoxOrient: 'vertical',
+  overflow: 'hidden',
+};
+
+export const PROVIDER_PRESET_CARD_TITLE_STYLE: CSSProperties = {
+  display: '-webkit-box',
+  WebkitLineClamp: 2,
+  WebkitBoxOrient: 'vertical',
+  overflow: 'hidden',
+};
--- a/frontend/wailsjs/go/aiservice/Service.d.ts
+++ b/frontend/wailsjs/go/aiservice/Service.d.ts
@@ -1,7 +1,6 @@
 // Cynhyrchwyd y ffeil hon yn awtomatig. PEIDIWCH Â MODIWL
 // This file is automatically generated. DO NOT EDIT
 import {ai} from '../models';
-import {context} from '../models';

 export function AIChatCancel(arg1:string):Promise<void>;

@@ -42,5 +41,3 @@ export function AISetContextLevel(arg1:string):Promise<void>;
 export function AISetSafetyLevel(arg1:string):Promise<void>;

 export function AITestProvider(arg1:ai.ProviderConfig):Promise<Record<string, any>>;
-
-export function Startup(arg1:context.Context):Promise<void>;
--- a/frontend/wailsjs/go/aiservice/Service.js
+++ b/frontend/wailsjs/go/aiservice/Service.js
@@ -81,7 +81,3 @@ export function AISetSafetyLevel(arg1) {
 export function AITestProvider(arg1) {
  return window['go']['aiservice']['Service']['AITestProvider'](arg1);
 }
-
-export function Startup(arg1) {
-  return window['go']['aiservice']['Service']['Startup'](arg1);
-}
--- a/frontend/wailsjs/go/app/App.d.ts
+++ b/frontend/wailsjs/go/app/App.d.ts
@@ -1,10 +1,8 @@
 // Cynhyrchwyd y ffeil hon yn awtomatig. PEIDIWCH Â MODIWL
 // This file is automatically generated. DO NOT EDIT
 import {connection} from '../models';
-import {time} from '../models';
 import {sync} from '../models';
 import {redis} from '../models';
-import {context} from '../models';

 export function ApplyChanges(arg1:connection.ConnectionConfig,arg2:string,arg3:string,arg4:connection.ChangeSet):Promise<connection.QueryResult>;

@@ -16,8 +14,6 @@ export function CheckDriverNetworkStatus():Promise<connection.QueryResult>;

 export function CheckForUpdates():Promise<connection.QueryResult>;

-export function CleanupStaleQueries(arg1:time.Duration):Promise<void>;
-
 export function ConfigureDriverRuntimeDirectory(arg1:string):Promise<connection.QueryResult>;

 export function ConfigureGlobalProxy(arg1:boolean,arg2:connection.ProxyConfig):Promise<connection.QueryResult>;
@@ -198,8 +194,6 @@ export function SetMacNativeWindowControls(arg1:boolean):Promise<void>;

 export function SetWindowTranslucency(arg1:number,arg2:number):Promise<void>;

-export function Startup(arg1:context.Context):Promise<void>;
-
 export function TestConnection(arg1:connection.ConnectionConfig):Promise<connection.QueryResult>;

 export function TruncateTables(arg1:connection.ConnectionConfig,arg2:string,arg3:Array<string>):Promise<connection.QueryResult>;
--- a/frontend/wailsjs/go/app/App.js
+++ b/frontend/wailsjs/go/app/App.js
@@ -22,10 +22,6 @@ export function CheckForUpdates() {
  return window['go']['app']['App']['CheckForUpdates']();
 }

-export function CleanupStaleQueries(arg1) {
-  return window['go']['app']['App']['CleanupStaleQueries'](arg1);
-}
-
 export function ConfigureDriverRuntimeDirectory(arg1) {
  return window['go']['app']['App']['ConfigureDriverRuntimeDirectory'](arg1);
 }
@@ -386,10 +382,6 @@ export function SetWindowTranslucency(arg1, arg2) {
  return window['go']['app']['App']['SetWindowTranslucency'](arg1, arg2);
 }

-export function Startup(arg1) {
-  return window['go']['app']['App']['Startup'](arg1);
-}
-
 export function TestConnection(arg1) {
  return window['go']['app']['App']['TestConnection'](arg1);
 }
--- a/internal/ai/provider/anthropic.go
+++ b/internal/ai/provider/anthropic.go
@@ -8,6 +8,7 @@ import (
 	"fmt"
 	"io"
 	"net/http"
+	"net/url"
 	"strings"

 	"GoNavi-Wails/internal/ai"
@@ -32,6 +33,25 @@ func normalizeAnthropicMessagesURL(baseURL string) string {
 	return url + "/v1/messages"
 }

+func IsDashScopeAnthropicCompatibleBaseURL(baseURL string) bool {
+	parsed, err := url.Parse(strings.TrimSpace(baseURL))
+	if err != nil {
+		return false
+	}
+	host := strings.ToLower(parsed.Hostname())
+	return host == "dashscope.aliyuncs.com" || host == "coding.dashscope.aliyuncs.com"
+}
+
+func ApplyAnthropicAuthHeaders(headers http.Header, baseURL string, apiKey string) {
+	headers.Set("x-api-key", apiKey)
+	if IsDashScopeAnthropicCompatibleBaseURL(baseURL) {
+		headers.Set("Authorization", "Bearer "+apiKey)
+		headers.Del("anthropic-version")
+		return
+	}
+	headers.Set("anthropic-version", anthropicAPIVersion)
+}
+
 // AnthropicProvider 实现 Anthropic Claude API 的 Provider
 type AnthropicProvider struct {
 	config  ai.ProviderConfig
@@ -446,8 +466,7 @@ func (p *AnthropicProvider) doRequest(ctx context.Context, body interface{}) (io
 	}

 	httpReq.Header.Set("Content-Type", "application/json")
-	httpReq.Header.Set("x-api-key", p.config.APIKey)
-	httpReq.Header.Set("anthropic-version", anthropicAPIVersion)
+	ApplyAnthropicAuthHeaders(httpReq.Header, p.baseURL, p.config.APIKey)

 	if strings.Contains(string(jsonBody), `"stream":true`) || strings.Contains(string(jsonBody), `"stream": true`) {
 		httpReq.Header.Set("Accept", "text/event-stream")
--- a/internal/ai/provider/anthropic_test.go
+++ b/internal/ai/provider/anthropic_test.go
@@ -1,6 +1,9 @@
 package provider

-import "testing"
+import (
+	"net/http"
+	"testing"
+)

 func TestNormalizeAnthropicMessagesURL_AppendsMessagesSuffix(t *testing.T) {
 	url := normalizeAnthropicMessagesURL("https://api.anthropic.com")
@@ -22,3 +25,33 @@ func TestNormalizeAnthropicMessagesURL_PreservesExplicitMessagesPath(t *testing.
 		t.Fatalf("expected explicit messages path to be preserved, got %q", url)
 	}
 }
+
+func TestApplyAnthropicAuthHeaders_UsesOfficialAnthropicHeadersForAnthropicAPI(t *testing.T) {
+	headers := http.Header{}
+	ApplyAnthropicAuthHeaders(headers, "https://api.anthropic.com", "sk-test")
+
+	if got := headers.Get("x-api-key"); got != "sk-test" {
+		t.Fatalf("expected x-api-key header, got %q", got)
+	}
+	if got := headers.Get("anthropic-version"); got != anthropicAPIVersion {
+		t.Fatalf("expected anthropic-version header, got %q", got)
+	}
+	if got := headers.Get("Authorization"); got != "" {
+		t.Fatalf("expected no authorization header for official anthropic, got %q", got)
+	}
+}
+
+func TestApplyAnthropicAuthHeaders_UsesBearerForDashScopeCompatibleAnthropic(t *testing.T) {
+	headers := http.Header{}
+	ApplyAnthropicAuthHeaders(headers, "https://coding.dashscope.aliyuncs.com/apps/anthropic", "sk-sp-test")
+
+	if got := headers.Get("Authorization"); got != "Bearer sk-sp-test" {
+		t.Fatalf("expected bearer authorization header, got %q", got)
+	}
+	if got := headers.Get("x-api-key"); got != "sk-sp-test" {
+		t.Fatalf("expected x-api-key header, got %q", got)
+	}
+	if got := headers.Get("anthropic-version"); got != "" {
+		t.Fatalf("expected no anthropic-version header for DashScope, got %q", got)
+	}
+}
--- a/internal/ai/provider/claude_cli.go
+++ b/internal/ai/provider/claude_cli.go
@@ -5,16 +5,20 @@ import (
 	"bytes"
 	"context"
 	"encoding/json"
+	"errors"
 	"fmt"
 	"os"
 	"os/exec"
 	"runtime"
 	"strings"
+	"time"

 	ai "GoNavi-Wails/internal/ai"
 )

 var claudeLookPath = exec.LookPath
+var claudeCommandContext = exec.CommandContext
+var claudeCLIRequestTimeout = 90 * time.Second

 // ClaudeCLIProvider 通过 Claude Code CLI 发送聊天请求
 // 适用于 anyrouter/newapi 等只支持 Claude Code 协议的代理服务
@@ -48,19 +52,25 @@ func (p *ClaudeCLIProvider) Chat(ctx context.Context, req ai.ChatRequest) (*ai.C
 		return nil, err
 	}

+	ctx, cancel := ensureClaudeCLITimeout(ctx, claudeCLIRequestTimeout)
+	defer cancel()
+
 	prompt := buildPrompt(req.Messages)
 	args := []string{"-p", prompt, "--output-format", "json", "--no-session-persistence"}
 	if p.config.Model != "" {
 		args = append(args, "--model", p.config.Model)
 	}

-	cmd := exec.CommandContext(ctx, "claude", args...)
+	cmd := claudeCommandContext(ctx, "claude", args...)
 	if err := p.setEnv(cmd); err != nil {
 		return nil, err
 	}

 	output, err := cmd.Output()
 	if err != nil {
+		if isClaudeCLITimeout(ctx, err) {
+			return nil, fmt.Errorf("claude CLI 执行超时（%s），当前 Base URL 或 API Key 可能没有返回有效响应", claudeCLIRequestTimeout)
+		}
 		if exitErr, ok := err.(*exec.ExitError); ok {
 			return nil, fmt.Errorf("claude CLI 执行失败: %s", string(exitErr.Stderr))
 		}
@@ -68,13 +78,14 @@ func (p *ClaudeCLIProvider) Chat(ctx context.Context, req ai.ChatRequest) (*ai.C
 	}

 	// 解析 JSON 输出
-	var result struct {
-		Result string `json:"result"`
-	}
+	var result cliStreamEvent
 	if err := json.Unmarshal(output, &result); err != nil {
 		// 如果 JSON 解析失败，直接返回原始文本
 		return &ai.ChatResponse{Content: strings.TrimSpace(string(output))}, nil
 	}
+	if errMsg, hasError := extractClaudeCLIEventError(result); hasError {
+		return nil, fmt.Errorf("claude CLI 返回错误: %s", errMsg)
+	}

 	return &ai.ChatResponse{Content: result.Result}, nil
 }
@@ -85,6 +96,9 @@ func (p *ClaudeCLIProvider) ChatStream(ctx context.Context, req ai.ChatRequest,
 		return err
 	}

+	ctx, cancel := ensureClaudeCLITimeout(ctx, claudeCLIRequestTimeout)
+	defer cancel()
+
 	prompt := buildPrompt(req.Messages)
 	args := []string{"-p", prompt, "--output-format", "stream-json", "--verbose", "--include-partial-messages", "--no-session-persistence"}
 	if p.config.Model != "" {
@@ -93,7 +107,7 @@ func (p *ClaudeCLIProvider) ChatStream(ctx context.Context, req ai.ChatRequest,

 	fmt.Printf("[ClaudeCLI DEBUG] Running: claude %v\n", args)

-	cmd := exec.CommandContext(ctx, "claude", args...)
+	cmd := claudeCommandContext(ctx, "claude", args...)
 	if err := p.setEnv(cmd); err != nil {
 		return err
 	}
@@ -137,7 +151,23 @@ func (p *ClaudeCLIProvider) ChatStream(ctx context.Context, req ai.ChatRequest,
 		}

 		switch event.Type {
+		case "system":
+			if isClaudeCLISystemRetryEvent(event) {
+				if errMsg, hasError := extractClaudeCLISystemRetryError(event); hasError {
+					callback(ai.StreamChunk{Error: errMsg, Done: true})
+					if cmd.Process != nil {
+						_ = cmd.Process.Kill()
+					}
+					_ = cmd.Wait()
+					return nil
+				}
+			}
 		case "assistant":
+			if errMsg, hasError := extractClaudeCLIEventError(event); hasError {
+				callback(ai.StreamChunk{Error: errMsg, Done: true})
+				_ = cmd.Wait()
+				return nil
+			}
 			// 助手消息开始或文本内容
 			if event.Message.Content != nil {
 				for _, block := range event.Message.Content {
@@ -156,12 +186,18 @@ func (p *ClaudeCLIProvider) ChatStream(ctx context.Context, req ai.ChatRequest,
 				callback(ai.StreamChunk{Content: event.Delta.Text})
 			}
 		case "result":
+			if errMsg, hasError := extractClaudeCLIEventError(event); hasError {
+				callback(ai.StreamChunk{Error: errMsg, Done: true})
+				_ = cmd.Wait()
+				return nil
+			}
 			// 最终结果事件 — 不发送 content（assistant 事件已包含），只标记完成
 			callback(ai.StreamChunk{Done: true})
 			_ = cmd.Wait()
 			return nil
 		case "error":
-			callback(ai.StreamChunk{Error: event.Error.Message, Done: true})
+			errMsg, _ := extractClaudeCLIEventError(event)
+			callback(ai.StreamChunk{Error: errMsg, Done: true})
 			_ = cmd.Wait()
 			return nil
 		}
@@ -171,6 +207,14 @@ func (p *ClaudeCLIProvider) ChatStream(ctx context.Context, req ai.ChatRequest,
 	stderrStr := strings.TrimSpace(stderrBuf.String())
 	fmt.Printf("[ClaudeCLI DEBUG] Process exited. stderr: %s\n", stderrStr)

+	if isClaudeCLITimeout(ctx, waitErr) {
+		callback(ai.StreamChunk{
+			Error: fmt.Sprintf("claude CLI 执行超时（%s），当前 Base URL 或 API Key 可能没有返回有效响应", claudeCLIRequestTimeout),
+			Done:  true,
+		})
+		return nil
+	}
+
 	if waitErr != nil {
 		errMsg := fmt.Sprintf("claude CLI 异常退出: %v", waitErr)
 		if stderrStr != "" {
@@ -184,6 +228,20 @@ func (p *ClaudeCLIProvider) ChatStream(ctx context.Context, req ai.ChatRequest,
 	return nil
 }

+func ensureClaudeCLITimeout(ctx context.Context, timeout time.Duration) (context.Context, context.CancelFunc) {
+	if _, hasDeadline := ctx.Deadline(); hasDeadline || timeout <= 0 {
+		return ctx, func() {}
+	}
+	return context.WithTimeout(ctx, timeout)
+}
+
+func isClaudeCLITimeout(ctx context.Context, err error) bool {
+	if err == nil {
+		return false
+	}
+	return errors.Is(ctx.Err(), context.DeadlineExceeded) || errors.Is(err, context.DeadlineExceeded)
+}
+
 // setEnv 设置 Claude CLI 的环境变量
 func (p *ClaudeCLIProvider) setEnv(cmd *exec.Cmd) error {
 	env, err := buildClaudeCLIEnv(p.config, cmd.Environ(), runtime.GOOS, claudeLookPath, fileExists)
@@ -200,6 +258,7 @@ func buildClaudeCLIEnv(config ai.ProviderConfig, baseEnv []string, goos string,
 		env = upsertEnv(env, "ANTHROPIC_BASE_URL", strings.TrimRight(config.BaseURL, "/"))
 	}
 	if config.APIKey != "" {
+		env = upsertEnv(env, "ANTHROPIC_AUTH_TOKEN", config.APIKey)
 		env = upsertEnv(env, "ANTHROPIC_API_KEY", config.APIKey)
 	}

@@ -354,8 +413,15 @@ func buildPrompt(messages []ai.Message) string {

 // cliStreamEvent Claude CLI stream-json 输出的事件结构
 type cliStreamEvent struct {
-	Type    string `json:"type"`
-	Message struct {
+	Type         string  `json:"type"`
+	Subtype      string  `json:"subtype,omitempty"`
+	IsError      bool    `json:"is_error,omitempty"`
+	Attempt      int     `json:"attempt,omitempty"`
+	MaxRetries   int     `json:"max_retries,omitempty"`
+	RetryDelayMS float64 `json:"retry_delay_ms,omitempty"`
+	ErrorStatus  int     `json:"error_status,omitempty"`
+	SessionID    string  `json:"session_id,omitempty"`
+	Message      struct {
 		Content []struct {
 			Type     string `json:"type"`
 			Text     string `json:"text"`
@@ -367,8 +433,79 @@ type cliStreamEvent struct {
 		Text     string `json:"text"`
 		Thinking string `json:"thinking"`
 	} `json:"delta,omitempty"`
-	Result string `json:"result,omitempty"`
-	Error  struct {
-		Message string `json:"message"`
-	} `json:"error,omitempty"`
+	Result string              `json:"result,omitempty"`
+	Error  cliStreamEventError `json:"error,omitempty"`
+}
+
+type cliStreamEventError struct {
+	Message string
+}
+
+func (e *cliStreamEventError) UnmarshalJSON(data []byte) error {
+	trimmed := strings.TrimSpace(string(data))
+	if trimmed == "" || trimmed == "null" {
+		e.Message = ""
+		return nil
+	}
+
+	var text string
+	if err := json.Unmarshal(data, &text); err == nil {
+		e.Message = strings.TrimSpace(text)
+		return nil
+	}
+
+	var payload struct {
+		Message string `json:"message"`
+	}
+	if err := json.Unmarshal(data, &payload); err != nil {
+		return err
+	}
+	e.Message = strings.TrimSpace(payload.Message)
+	return nil
+}
+
+func extractClaudeCLIEventError(event cliStreamEvent) (string, bool) {
+	if event.Type != "error" && !event.IsError {
+		return "", false
+	}
+
+	if msg := strings.TrimSpace(event.Result); msg != "" {
+		return msg, true
+	}
+
+	for _, block := range event.Message.Content {
+		if block.Type == "text" && strings.TrimSpace(block.Text) != "" {
+			return strings.TrimSpace(block.Text), true
+		}
+	}
+
+	if msg := strings.TrimSpace(event.Error.Message); msg != "" {
+		return msg, true
+	}
+
+	return "claude CLI 返回未知错误", true
+}
+
+func isClaudeCLISystemRetryEvent(event cliStreamEvent) bool {
+	return event.Type == "system" && event.Subtype == "api_retry"
+}
+
+func extractClaudeCLISystemRetryError(event cliStreamEvent) (string, bool) {
+	if !isClaudeCLISystemRetryEvent(event) {
+		return "", false
+	}
+
+	errText := strings.TrimSpace(event.Error.Message)
+	if event.ErrorStatus != 401 && event.ErrorStatus != 403 && !strings.EqualFold(errText, "authentication_failed") {
+		return "", false
+	}
+
+	if errText == "" {
+		errText = "authentication_failed"
+	}
+
+	if event.ErrorStatus > 0 {
+		return fmt.Sprintf("claude CLI 鉴权失败 (HTTP %d): %s", event.ErrorStatus, errText), true
+	}
+	return fmt.Sprintf("claude CLI 鉴权失败: %s", errText), true
 }
--- a/internal/ai/provider/claude_cli_test.go
+++ b/internal/ai/provider/claude_cli_test.go
@@ -1,9 +1,13 @@
 package provider

 import (
+	"context"
 	"errors"
+	"os"
+	"path/filepath"
 	"strings"
 	"testing"
+	"time"

 	"GoNavi-Wails/internal/ai"
 )
@@ -26,6 +30,9 @@ func TestBuildClaudeCLIEnv_IncludesAnthropicProxyEnv(t *testing.T) {
 	if got := envValue(env, "ANTHROPIC_API_KEY"); got != "sk-test" {
 		t.Fatalf("expected api key in env, got %q", got)
 	}
+	if got := envValue(env, "ANTHROPIC_AUTH_TOKEN"); got != "sk-test" {
+		t.Fatalf("expected auth token in env, got %q", got)
+	}
 }

 func TestBuildClaudeCLIEnv_UsesDetectedGitBashOnWindows(t *testing.T) {
@@ -67,3 +74,281 @@ func TestBuildClaudeCLIEnv_ReturnsActionableErrorWhenGitBashMissingOnWindows(t *
 		t.Fatalf("expected env var hint, got %v", err)
 	}
 }
+
+func TestClaudeCLIProvider_ChatTimesOutWhenCommandDoesNotFinish(t *testing.T) {
+	fakeClaude := writeFakeClaudeScript(t, "#!/bin/sh\nsleep 5\n")
+	restore := overrideClaudeCLIForTest(t, fakeClaude)
+	defer restore()
+
+	originalRequestTimeout := claudeCLIRequestTimeout
+	claudeCLIRequestTimeout = 200 * time.Millisecond
+	defer func() {
+		claudeCLIRequestTimeout = originalRequestTimeout
+	}()
+
+	provider, err := NewClaudeCLIProvider(ai.ProviderConfig{
+		BaseURL: "https://coding.dashscope.aliyuncs.com/apps/anthropic",
+		APIKey:  "sk-test",
+		Model:   "qwen3.5-plus",
+	})
+	if err != nil {
+		t.Fatalf("unexpected provider error: %v", err)
+	}
+
+	start := time.Now()
+	_, err = provider.Chat(context.Background(), ai.ChatRequest{
+		Messages: []ai.Message{{Role: "user", Content: "ping"}},
+	})
+	if err == nil {
+		t.Fatal("expected chat timeout error")
+	}
+	if !strings.Contains(err.Error(), "执行超时") {
+		t.Fatalf("expected timeout error, got %v", err)
+	}
+	if time.Since(start) < 200*time.Millisecond {
+		t.Fatalf("expected timeout path to wait for configured deadline, took %s", time.Since(start))
+	}
+}
+
+func TestClaudeCLIProvider_ChatStreamUsesRequestTimeoutWhenNoMeaningfulResponseArrives(t *testing.T) {
+	fakeClaude := writeFakeClaudeScript(t, "#!/bin/sh\necho '{\"type\":\"system\",\"subtype\":\"init\"}'\nexec sleep 5\n")
+	restore := overrideClaudeCLIForTest(t, fakeClaude)
+	defer restore()
+
+	originalRequestTimeout := claudeCLIRequestTimeout
+	claudeCLIRequestTimeout = 200 * time.Millisecond
+	defer func() {
+		claudeCLIRequestTimeout = originalRequestTimeout
+	}()
+
+	provider, err := NewClaudeCLIProvider(ai.ProviderConfig{
+		BaseURL: "https://coding.dashscope.aliyuncs.com/apps/anthropic",
+		APIKey:  "sk-test",
+		Model:   "qwen3.5-plus",
+	})
+	if err != nil {
+		t.Fatalf("unexpected provider error: %v", err)
+	}
+
+	var chunks []ai.StreamChunk
+	err = provider.ChatStream(context.Background(), ai.ChatRequest{
+		Messages: []ai.Message{{Role: "user", Content: "ping"}},
+	}, func(chunk ai.StreamChunk) {
+		chunks = append(chunks, chunk)
+	})
+	if err != nil {
+		t.Fatalf("expected stream provider to report timeout via callback, got %v", err)
+	}
+	if len(chunks) == 0 {
+		t.Fatal("expected timeout chunk")
+	}
+	lastChunk := chunks[len(chunks)-1]
+	if !lastChunk.Done {
+		t.Fatalf("expected timeout chunk to terminate stream, got %#v", lastChunk)
+	}
+	if !strings.Contains(lastChunk.Error, "执行超时") {
+		t.Fatalf("expected request timeout message, got %#v", lastChunk)
+	}
+}
+
+func TestClaudeCLIProvider_ChatStreamAllowsDelayedMeaningfulResponse(t *testing.T) {
+	fakeClaude := writeFakeClaudeScript(t, "#!/bin/sh\necho '{\"type\":\"system\",\"subtype\":\"init\"}'\nsleep 0.2\necho '{\"type\":\"assistant\",\"message\":{\"content\":[{\"type\":\"text\",\"text\":\"OK\"}]}}'\necho '{\"type\":\"result\",\"subtype\":\"success\",\"is_error\":false,\"result\":\"OK\"}'\n")
+	restore := overrideClaudeCLIForTest(t, fakeClaude)
+	defer restore()
+
+	originalRequestTimeout := claudeCLIRequestTimeout
+	claudeCLIRequestTimeout = 1 * time.Second
+	defer func() {
+		claudeCLIRequestTimeout = originalRequestTimeout
+	}()
+
+	provider, err := NewClaudeCLIProvider(ai.ProviderConfig{
+		BaseURL: "https://coding.dashscope.aliyuncs.com/apps/anthropic",
+		APIKey:  "sk-test",
+		Model:   "qwen3.5-plus",
+	})
+	if err != nil {
+		t.Fatalf("unexpected provider error: %v", err)
+	}
+
+	var chunks []ai.StreamChunk
+	err = provider.ChatStream(context.Background(), ai.ChatRequest{
+		Messages: []ai.Message{{Role: "user", Content: "ping"}},
+	}, func(chunk ai.StreamChunk) {
+		chunks = append(chunks, chunk)
+	})
+	if err != nil {
+		t.Fatalf("expected delayed response to complete via callback, got %v", err)
+	}
+	if len(chunks) == 0 {
+		t.Fatal("expected delayed response chunks")
+	}
+	if chunks[0].Content != "OK" {
+		t.Fatalf("expected delayed content chunk, got %#v", chunks)
+	}
+	if !chunks[len(chunks)-1].Done {
+		t.Fatalf("expected terminal done chunk, got %#v", chunks[len(chunks)-1])
+	}
+}
+
+func TestClaudeCLIProvider_ChatReturnsErrorWhenJSONResponseIsError(t *testing.T) {
+	fakeClaude := writeFakeClaudeScript(t, "#!/bin/sh\necho '{\"type\":\"result\",\"subtype\":\"success\",\"is_error\":true,\"result\":\"API Error: Unable to connect to API (ECONNRESET)\",\"error\":\"unknown\"}'\n")
+	restore := overrideClaudeCLIForTest(t, fakeClaude)
+	defer restore()
+
+	provider, err := NewClaudeCLIProvider(ai.ProviderConfig{
+		BaseURL: "https://coding.dashscope.aliyuncs.com/apps/anthropic",
+		APIKey:  "sk-test",
+		Model:   "qwen3.5-plus",
+	})
+	if err != nil {
+		t.Fatalf("unexpected provider error: %v", err)
+	}
+
+	_, err = provider.Chat(context.Background(), ai.ChatRequest{
+		Messages: []ai.Message{{Role: "user", Content: "ping"}},
+	})
+	if err == nil {
+		t.Fatal("expected chat error when CLI JSON marks request as failed")
+	}
+	if !strings.Contains(err.Error(), "Unable to connect to API") {
+		t.Fatalf("expected upstream API error, got %v", err)
+	}
+}
+
+func TestClaudeCLIProvider_ChatStreamReportsAssistantErrorEvent(t *testing.T) {
+	fakeClaude := writeFakeClaudeScript(t, "#!/bin/sh\necho '{\"type\":\"assistant\",\"is_error\":true,\"message\":{\"content\":[{\"type\":\"text\",\"text\":\"API Error: Unable to connect to API (ECONNRESET)\"}]},\"error\":\"unknown\"}'\n")
+	restore := overrideClaudeCLIForTest(t, fakeClaude)
+	defer restore()
+
+	provider, err := NewClaudeCLIProvider(ai.ProviderConfig{
+		BaseURL: "https://coding.dashscope.aliyuncs.com/apps/anthropic",
+		APIKey:  "sk-test",
+		Model:   "qwen3.5-plus",
+	})
+	if err != nil {
+		t.Fatalf("unexpected provider error: %v", err)
+	}
+
+	var chunks []ai.StreamChunk
+	err = provider.ChatStream(context.Background(), ai.ChatRequest{
+		Messages: []ai.Message{{Role: "user", Content: "ping"}},
+	}, func(chunk ai.StreamChunk) {
+		chunks = append(chunks, chunk)
+	})
+	if err != nil {
+		t.Fatalf("expected stream provider to report error via callback, got %v", err)
+	}
+	if len(chunks) != 1 {
+		t.Fatalf("expected a single terminal error chunk, got %#v", chunks)
+	}
+	if chunks[0].Content != "" {
+		t.Fatalf("expected assistant error event to avoid content output, got %#v", chunks[0])
+	}
+	if !chunks[0].Done || !strings.Contains(chunks[0].Error, "Unable to connect to API") {
+		t.Fatalf("expected upstream API error chunk, got %#v", chunks[0])
+	}
+}
+
+func TestClaudeCLIProvider_ChatStreamReportsResultErrorEvent(t *testing.T) {
+	fakeClaude := writeFakeClaudeScript(t, "#!/bin/sh\necho '{\"type\":\"result\",\"subtype\":\"success\",\"is_error\":true,\"result\":\"API Error: Unable to connect to API (ECONNRESET)\",\"error\":\"unknown\"}'\n")
+	restore := overrideClaudeCLIForTest(t, fakeClaude)
+	defer restore()
+
+	provider, err := NewClaudeCLIProvider(ai.ProviderConfig{
+		BaseURL: "https://coding.dashscope.aliyuncs.com/apps/anthropic",
+		APIKey:  "sk-test",
+		Model:   "qwen3.5-plus",
+	})
+	if err != nil {
+		t.Fatalf("unexpected provider error: %v", err)
+	}
+
+	var chunks []ai.StreamChunk
+	err = provider.ChatStream(context.Background(), ai.ChatRequest{
+		Messages: []ai.Message{{Role: "user", Content: "ping"}},
+	}, func(chunk ai.StreamChunk) {
+		chunks = append(chunks, chunk)
+	})
+	if err != nil {
+		t.Fatalf("expected stream provider to report error via callback, got %v", err)
+	}
+	if len(chunks) != 1 {
+		t.Fatalf("expected a single terminal error chunk, got %#v", chunks)
+	}
+	if chunks[0].Content != "" {
+		t.Fatalf("expected result error event to avoid content output, got %#v", chunks[0])
+	}
+	if !chunks[0].Done || !strings.Contains(chunks[0].Error, "Unable to connect to API") {
+		t.Fatalf("expected upstream API error chunk, got %#v", chunks[0])
+	}
+}
+
+func TestClaudeCLIProvider_ChatStreamReportsApiRetryAuthenticationFailure(t *testing.T) {
+	fakeClaude := writeFakeClaudeScript(t, "#!/bin/sh\necho '{\"type\":\"system\",\"subtype\":\"api_retry\",\"attempt\":1,\"max_retries\":10,\"retry_delay_ms\":536.11,\"error_status\":401,\"error\":\"authentication_failed\",\"session_id\":\"retry-1\"}'\nexec sleep 5\n")
+	restore := overrideClaudeCLIForTest(t, fakeClaude)
+	defer restore()
+
+	provider, err := NewClaudeCLIProvider(ai.ProviderConfig{
+		BaseURL: "https://coding.dashscope.aliyuncs.com/apps/anthropic",
+		APIKey:  "sk-test",
+		Model:   "qwen3.5-plus",
+	})
+	if err != nil {
+		t.Fatalf("unexpected provider error: %v", err)
+	}
+
+	var chunks []ai.StreamChunk
+	err = provider.ChatStream(context.Background(), ai.ChatRequest{
+		Messages: []ai.Message{{Role: "user", Content: "ping"}},
+	}, func(chunk ai.StreamChunk) {
+		chunks = append(chunks, chunk)
+	})
+	if err != nil {
+		t.Fatalf("expected stream provider to report authentication error via callback, got %v", err)
+	}
+	if len(chunks) != 1 {
+		t.Fatalf("expected a single terminal error chunk, got %#v", chunks)
+	}
+	if !chunks[0].Done {
+		t.Fatalf("expected terminal error chunk, got %#v", chunks[0])
+	}
+	if strings.Contains(chunks[0].Error, "未收到模型响应") {
+		t.Fatalf("expected auth failure instead of startup timeout, got %#v", chunks[0])
+	}
+	if !strings.Contains(chunks[0].Error, "401") || !strings.Contains(chunks[0].Error, "authentication_failed") {
+		t.Fatalf("expected auth retry error details, got %#v", chunks[0])
+	}
+}
+
+func writeFakeClaudeScript(t *testing.T, content string) string {
+	t.Helper()
+	dir := t.TempDir()
+	path := filepath.Join(dir, "claude")
+	if err := os.WriteFile(path, []byte(content), 0o755); err != nil {
+		t.Fatalf("failed to write fake claude script: %v", err)
+	}
+	return path
+}
+
+func overrideClaudeCLIForTest(t *testing.T, fakeClaudePath string) func() {
+	t.Helper()
+
+	originalLookPath := claudeLookPath
+	claudeLookPath = func(name string) (string, error) {
+		if name == "claude" {
+			return fakeClaudePath, nil
+		}
+		return originalLookPath(name)
+	}
+
+	originalPath := os.Getenv("PATH")
+	if err := os.Setenv("PATH", filepath.Dir(fakeClaudePath)+string(os.PathListSeparator)+originalPath); err != nil {
+		t.Fatalf("failed to override PATH: %v", err)
+	}
+
+	return func() {
+		claudeLookPath = originalLookPath
+		_ = os.Setenv("PATH", originalPath)
+	}
+}
--- a/internal/ai/provider/helper.go
+++ b/internal/ai/provider/helper.go
@@ -2,9 +2,13 @@ package provider

 import (
 	"fmt"
+	"net/url"
+	"regexp"
 	"strings"
 )

+var openAICompatibleVersionSuffixPattern = regexp.MustCompile(`(?i)(^|/)v\d+$`)
+
 // ParseDataURI 解析前端传递的 Data URI，返回 mimeType 和去掉前缀的 rawBase64
 func ParseDataURI(dataURI string) (mimeType, rawBase64 string, err error) {
 	if !strings.HasPrefix(dataURI, "data:") {
@@ -24,3 +28,70 @@ func ParseDataURI(dataURI string) (mimeType, rawBase64 string, err error) {
 	rawBase64 = parts[1]
 	return mimeType, rawBase64, nil
 }
+
+// NormalizeOpenAICompatibleBaseURL 统一归一化 OpenAI 兼容服务的 base URL。
+func NormalizeOpenAICompatibleBaseURL(raw string) string {
+	trimmed := strings.TrimSpace(raw)
+	if trimmed == "" {
+		return defaultOpenAIBaseURL
+	}
+
+	parsed, err := url.Parse(trimmed)
+	if err != nil || parsed.Scheme == "" || parsed.Host == "" {
+		return normalizeOpenAICompatibleBaseURLString(trimmed)
+	}
+
+	parsed.RawQuery = ""
+	parsed.Fragment = ""
+	parsed.Path = normalizeOpenAICompatiblePath(parsed.Path)
+	return strings.TrimRight(parsed.String(), "/")
+}
+
+// ResolveOpenAICompatibleEndpoint 基于归一化 base URL 拼接 OpenAI 兼容接口路径。
+func ResolveOpenAICompatibleEndpoint(baseURL string, endpoint string) string {
+	normalizedBaseURL := NormalizeOpenAICompatibleBaseURL(baseURL)
+	normalizedEndpoint := strings.TrimLeft(strings.TrimSpace(endpoint), "/")
+	if normalizedEndpoint == "" {
+		return normalizedBaseURL
+	}
+	return normalizedBaseURL + "/" + normalizedEndpoint
+}
+
+func normalizeOpenAICompatibleBaseURLString(raw string) string {
+	normalized := strings.TrimRight(strings.TrimSpace(raw), "/")
+	if normalized == "" {
+		return defaultOpenAIBaseURL
+	}
+
+	lower := strings.ToLower(normalized)
+	switch {
+	case strings.HasSuffix(lower, "/chat/completions"):
+		normalized = normalized[:len(normalized)-len("/chat/completions")]
+	case strings.HasSuffix(lower, "/models"):
+		normalized = normalized[:len(normalized)-len("/models")]
+	}
+	normalized = strings.TrimRight(normalized, "/")
+	if openAICompatibleVersionSuffixPattern.MatchString(normalized) {
+		return normalized
+	}
+	return normalized + "/v1"
+}
+
+func normalizeOpenAICompatiblePath(path string) string {
+	normalized := strings.TrimRight(strings.TrimSpace(path), "/")
+	lower := strings.ToLower(normalized)
+	switch {
+	case strings.HasSuffix(lower, "/chat/completions"):
+		normalized = normalized[:len(normalized)-len("/chat/completions")]
+	case strings.HasSuffix(lower, "/models"):
+		normalized = normalized[:len(normalized)-len("/models")]
+	}
+	normalized = strings.TrimRight(normalized, "/")
+	if openAICompatibleVersionSuffixPattern.MatchString(normalized) {
+		return normalized
+	}
+	if normalized == "" {
+		return "/v1"
+	}
+	return normalized + "/v1"
+}
--- a/internal/ai/provider/openai.go
+++ b/internal/ai/provider/openai.go
@@ -30,14 +30,7 @@ type OpenAIProvider struct {

 // NewOpenAIProvider 创建 OpenAI Provider 实例
 func NewOpenAIProvider(config ai.ProviderConfig) (Provider, error) {
-	baseURL := strings.TrimRight(strings.TrimSpace(config.BaseURL), "/")
-	if baseURL == "" {
-		baseURL = defaultOpenAIBaseURL
-	}
-	// 确保 baseURL 包含 /v1 路径（兼容用户只填域名的情况，如 https://anyrouter.top）
-	if !strings.HasSuffix(baseURL, "/v1") && !strings.Contains(baseURL, "/v1/") {
-		baseURL = baseURL + "/v1"
-	}
+	baseURL := NormalizeOpenAICompatibleBaseURL(config.BaseURL)
 	model := strings.TrimSpace(config.Model)
 	if model == "" {
 		return nil, fmt.Errorf("模型 ID 不能为空，请在设置中选择或输入模型")
@@ -315,7 +308,7 @@ func (p *OpenAIProvider) ChatStream(ctx context.Context, req ai.ChatRequest, cal
 		}
 		if len(chunk.Choices) > 0 {
 			choice := chunk.Choices[0]
-			
+
 			// Handle ToolCalls delta
 			if len(choice.Delta.ToolCalls) > 0 {
 				receivedContent = true
@@ -383,10 +376,7 @@ func (p *OpenAIProvider) doRequest(ctx context.Context, body interface{}) (io.Re
 		return nil, fmt.Errorf("序列化请求失败: %w", err)
 	}

-	url := p.baseURL + "/chat/completions"
-
-
-
+	url := ResolveOpenAICompatibleEndpoint(p.baseURL, "chat/completions")
 	httpReq, err := http.NewRequestWithContext(ctx, http.MethodPost, url, bytes.NewReader(jsonBody))
 	if err != nil {
 		return nil, fmt.Errorf("创建 HTTP 请求失败: %w", err)
--- a/internal/ai/provider/openai_test.go
+++ b/internal/ai/provider/openai_test.go
@@ -5,6 +5,76 @@ import (
 	"testing"
 )

+func TestNormalizeOpenAICompatibleBaseURL(t *testing.T) {
+	tests := []struct {
+		name string
+		raw  string
+		want string
+	}{
+		{
+			name: "empty uses default openai base url",
+			raw:  "",
+			want: "https://api.openai.com/v1",
+		},
+		{
+			name: "domain only appends v1",
+			raw:  "https://api.openai.com",
+			want: "https://api.openai.com/v1",
+		},
+		{
+			name: "keeps existing v1 suffix",
+			raw:  "https://api.deepseek.com/v1",
+			want: "https://api.deepseek.com/v1",
+		},
+		{
+			name: "keeps dashscope compatible mode path",
+			raw:  "https://dashscope.aliyuncs.com/compatible-mode/v1",
+			want: "https://dashscope.aliyuncs.com/compatible-mode/v1",
+		},
+		{
+			name: "keeps zhipu v4 path",
+			raw:  "https://open.bigmodel.cn/api/paas/v4",
+			want: "https://open.bigmodel.cn/api/paas/v4",
+		},
+		{
+			name: "keeps volcengine ark v3 path",
+			raw:  "https://ark.cn-beijing.volces.com/api/v3",
+			want: "https://ark.cn-beijing.volces.com/api/v3",
+		},
+		{
+			name: "keeps volcengine coding plan v3 path",
+			raw:  "https://ark.cn-beijing.volces.com/api/coding/v3",
+			want: "https://ark.cn-beijing.volces.com/api/coding/v3",
+		},
+		{
+			name: "strips chat completions suffix before normalizing",
+			raw:  "https://api.openai.com/v1/chat/completions",
+			want: "https://api.openai.com/v1",
+		},
+		{
+			name: "strips models suffix before normalizing",
+			raw:  "https://ark.cn-beijing.volces.com/api/coding/v3/models",
+			want: "https://ark.cn-beijing.volces.com/api/coding/v3",
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			if got := NormalizeOpenAICompatibleBaseURL(tt.raw); got != tt.want {
+				t.Fatalf("expected normalized base url %q, got %q", tt.want, got)
+			}
+		})
+	}
+}
+
+func TestResolveOpenAICompatibleEndpoint(t *testing.T) {
+	got := ResolveOpenAICompatibleEndpoint("https://ark.cn-beijing.volces.com/api/coding/v3/models", "chat/completions")
+	want := "https://ark.cn-beijing.volces.com/api/coding/v3/chat/completions"
+	if got != want {
+		t.Fatalf("expected endpoint %q, got %q", want, got)
+	}
+}
+
 func TestOpenAIProvider_Validate_MissingAPIKey(t *testing.T) {
 	p, err := NewOpenAIProvider(ai.ProviderConfig{Type: "openai", Model: "gpt-4o"})
 	if err != nil {
--- a/internal/ai/service/service.go
+++ b/internal/ai/service/service.go
@@ -6,6 +6,7 @@ import (
 	"fmt"
 	"io"
 	"net/http"
+	"net/url"
 	"os"
 	"path/filepath"
 	"strings"
@@ -31,7 +32,7 @@ type Service struct {
 	safetyLevel    ai.SQLPermissionLevel
 	contextLevel   ai.ContextLevel
 	guard          *safety.Guard
-	configDir      string // 配置存储目录
+	configDir      string                        // 配置存储目录
 	cancelFuncs    map[string]context.CancelFunc // 记录每个 session 的 context 取消函数
 }

@@ -45,6 +46,55 @@ var miniMaxAnthropicModels = []string{
 	"MiniMax-M2",
 }

+var dashScopeCodingPlanModels = []string{
+	"qwen3.5-plus",
+	"kimi-k2.5",
+	"glm-5",
+	"MiniMax-M2.5",
+	"qwen3-max-2026-01-23",
+	"qwen3-coder-next",
+	"qwen3-coder-plus",
+	"glm-4.7",
+}
+
+const dashScopeCodingPlanAnthropicBaseURL = "https://coding.dashscope.aliyuncs.com/apps/anthropic"
+
+var volcengineCodingPlanAllowedExactModels = []string{
+	"auto",
+}
+
+var volcengineCodingPlanAllowedModelFamilies = []string{
+	"doubao-seed-2.0-code",
+	"doubao-seed-2.0-pro",
+	"doubao-seed-2.0-lite",
+	"doubao-seed-code",
+	"minimax-m2.5",
+	"glm-4.7",
+	"deepseek-v3.2",
+	"kimi-k2",
+}
+
+const volcengineCodingPlanEmptyModelsError = `当前接口未返回可用的火山 Coding Plan 模型，请检查账号权限或切换到"火山方舟"供应商`
+
+var claudeCLIHealthCheckFunc = func(config ai.ProviderConfig) error {
+	ctx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
+	defer cancel()
+
+	cliProvider, err := provider.NewProvider(config)
+	if err != nil {
+		return err
+	}
+
+	_, err = cliProvider.Chat(ctx, ai.ChatRequest{
+		Messages: []ai.Message{
+			{Role: "user", Content: "ping"},
+		},
+		MaxTokens:   1,
+		Temperature: 0,
+	})
+	return err
+}
+
 // NewService 创建 AI Service 实例
 func NewService() *Service {
 	return &Service{
@@ -56,8 +106,13 @@ func NewService() *Service {
 	}
 }

-// Startup Wails 生命周期回调
-func (s *Service) Startup(ctx context.Context) {
+// InitializeLifecycle attaches runtime context without exposing lifecycle internals to Wails bindings.
+func InitializeLifecycle(s *Service, ctx context.Context) {
+	s.startup(ctx)
+}
+
+// startup Wails 生命周期回调
+func (s *Service) startup(ctx context.Context) {
 	s.ctx = ctx
 	s.configDir = resolveConfigDir()
 	s.loadConfig()
@@ -173,6 +228,12 @@ func (s *Service) AITestProvider(config ai.ProviderConfig) map[string]interface{
 				err = fmt.Errorf("上游服务器内部错误 (HTTP %d)", resp.StatusCode)
 			}
 		}
+	case "claude-cli":
+		testConfig := config
+		if strings.TrimSpace(testConfig.Model) == "" && isDashScopeCodingPlanProvider(testConfig) && len(dashScopeCodingPlanModels) > 0 {
+			testConfig.Model = dashScopeCodingPlanModels[0]
+		}
+		err = claudeCLIHealthCheckFunc(testConfig)
 	default:
 		if baseURL != "" {
 			req, _ := http.NewRequest("GET", baseURL, nil)
@@ -219,18 +280,104 @@ func isMoonshotAnthropicProvider(config ai.ProviderConfig) bool {
 	return strings.Contains(baseURL, "api.moonshot.cn")
 }

+func parseProviderBaseURL(raw string) (string, string) {
+	parsed, err := url.Parse(strings.TrimSpace(raw))
+	if err != nil {
+		return "", ""
+	}
+	return strings.ToLower(parsed.Hostname()), strings.TrimRight(strings.ToLower(parsed.Path), "/")
+}
+
+func isDashScopeBailianAnthropicProvider(config ai.ProviderConfig) bool {
+	if normalizedProviderType(config) != "anthropic" {
+		return false
+	}
+	host, path := parseProviderBaseURL(config.BaseURL)
+	return host == "dashscope.aliyuncs.com" && strings.HasPrefix(path, "/apps/anthropic")
+}
+
+func isDashScopeCodingPlanAnthropicProvider(config ai.ProviderConfig) bool {
+	if normalizedProviderType(config) != "anthropic" {
+		return false
+	}
+	return isDashScopeCodingPlanProvider(config)
+}
+
+func isDashScopeCodingPlanProvider(config ai.ProviderConfig) bool {
+	host, path := parseProviderBaseURL(config.BaseURL)
+	return host == "coding.dashscope.aliyuncs.com" && (strings.HasPrefix(path, "/apps/anthropic") || strings.HasPrefix(path, "/v1"))
+}
+
+func isVolcengineCodingPlanProvider(config ai.ProviderConfig) bool {
+	if normalizedProviderType(config) != "openai" {
+		return false
+	}
+	host, path := parseProviderBaseURL(provider.NormalizeOpenAICompatibleBaseURL(config.BaseURL))
+	return host == "ark.cn-beijing.volces.com" && path == "/api/coding/v3"
+}
+
+func filterVolcengineCodingPlanModels(models []string) []string {
+	filtered := make([]string, 0, len(models))
+	for _, model := range models {
+		lowerModel := strings.ToLower(strings.TrimSpace(model))
+		matched := false
+		for _, exactModel := range volcengineCodingPlanAllowedExactModels {
+			if lowerModel == exactModel {
+				filtered = append(filtered, model)
+				matched = true
+				break
+			}
+		}
+		if matched {
+			continue
+		}
+		for _, family := range volcengineCodingPlanAllowedModelFamilies {
+			if strings.Contains(lowerModel, family) {
+				filtered = append(filtered, model)
+				break
+			}
+		}
+	}
+	return filtered
+}
+
+func filterFetchedModelsForProvider(config ai.ProviderConfig, models []string) ([]string, error) {
+	if !isVolcengineCodingPlanProvider(config) {
+		return models, nil
+	}
+	filtered := filterVolcengineCodingPlanModels(models)
+	if len(filtered) == 0 {
+		return nil, fmt.Errorf(volcengineCodingPlanEmptyModelsError)
+	}
+	return filtered, nil
+}
+
 func defaultStaticModelsForProvider(config ai.ProviderConfig) []string {
 	if isMiniMaxAnthropicProvider(config) {
 		return append([]string(nil), miniMaxAnthropicModels...)
 	}
+	if isDashScopeCodingPlanProvider(config) {
+		return append([]string(nil), dashScopeCodingPlanModels...)
+	}
 	return nil
 }

 func normalizeProviderConfig(config ai.ProviderConfig) ai.ProviderConfig {
-	staticModels := defaultStaticModelsForProvider(config)
-	if len(staticModels) > 0 && len(config.Models) == 0 {
-		config.Models = staticModels
+	switch {
+	case isDashScopeBailianAnthropicProvider(config):
+		config.Models = nil
+	case isDashScopeCodingPlanProvider(config):
+		config.Type = "custom"
+		config.APIFormat = "claude-cli"
+		config.BaseURL = dashScopeCodingPlanAnthropicBaseURL
+		config.Models = append([]string(nil), dashScopeCodingPlanModels...)
+	default:
+		staticModels := defaultStaticModelsForProvider(config)
+		if len(staticModels) > 0 && len(config.Models) == 0 {
+			config.Models = staticModels
+		}
 	}
+
 	model := strings.TrimSpace(config.Model)
 	if isMiniMaxAnthropicProvider(config) && (model == "" || strings.HasPrefix(strings.ToLower(model), "minimax-text-")) {
 		config.Model = miniMaxAnthropicModels[0]
@@ -248,6 +395,9 @@ func resolveModelsURL(config ai.ProviderConfig) string {
 		if isMoonshotAnthropicProvider(config) {
 			return "https://api.moonshot.cn/v1/models"
 		}
+		if isDashScopeBailianAnthropicProvider(config) {
+			return "https://dashscope.aliyuncs.com/compatible-mode/v1/models"
+		}
 		if baseURL == "" {
 			baseURL = "https://api.anthropic.com"
 		}
@@ -263,13 +413,7 @@ func resolveModelsURL(config ai.ProviderConfig) string {
 	case "openai":
 		fallthrough
 	default:
-		if baseURL == "" {
-			baseURL = "https://api.openai.com/v1"
-		}
-		if !strings.HasSuffix(baseURL, "/v1") && !strings.Contains(baseURL, "/v1/") {
-			baseURL = baseURL + "/v1"
-		}
-		return baseURL + "/models"
+		return provider.ResolveOpenAICompatibleEndpoint(baseURL, "models")
 	}
 }

@@ -283,9 +427,11 @@ func newModelsRequest(config ai.ProviderConfig) (*http.Request, error) {

 	switch normalizedProviderType(config) {
 	case "anthropic":
-		req.Header.Set("x-api-key", config.APIKey)
-		req.Header.Set("anthropic-version", "2023-06-01")
-		req.Header.Set("Authorization", "Bearer "+config.APIKey)
+		if isDashScopeBailianAnthropicProvider(config) {
+			req.Header.Set("Authorization", "Bearer "+config.APIKey)
+		} else {
+			provider.ApplyAnthropicAuthHeaders(req.Header, config.BaseURL, config.APIKey)
+		}
 	case "gemini":
 		// Gemini 使用 query string 传递 key，无需额外鉴权头
 	default:
@@ -315,33 +461,36 @@ func resolveAnthropicMessagesURL(baseURL string) string {

 func newProviderHealthCheckRequest(config ai.ProviderConfig) (*http.Request, error) {
 	config = normalizeProviderConfig(config)
-	if isMiniMaxAnthropicProvider(config) {
-		body := map[string]interface{}{
-			"model":      config.Model,
-			"max_tokens": 1,
-			"messages": []map[string]string{
-				{"role": "user", "content": "ping"},
-			},
-		}
-		bodyBytes, err := json.Marshal(body)
-		if err != nil {
-			return nil, fmt.Errorf("序列化请求失败: %w", err)
-		}
-		req, err := http.NewRequest("POST", resolveAnthropicMessagesURL(config.BaseURL), strings.NewReader(string(bodyBytes)))
-		if err != nil {
-			return nil, fmt.Errorf("创建请求失败: %w", err)
-		}
-		req.Header.Set("Content-Type", "application/json")
-		req.Header.Set("x-api-key", config.APIKey)
-		req.Header.Set("anthropic-version", "2023-06-01")
-		for k, v := range config.Headers {
-			req.Header.Set(k, v)
-		}
-		return req, nil
+	if isMiniMaxAnthropicProvider(config) || isDashScopeBailianAnthropicProvider(config) || isDashScopeCodingPlanAnthropicProvider(config) {
+		return newAnthropicMessagesHealthCheckRequest(config)
 	}
 	return newModelsRequest(config)
 }

+func newAnthropicMessagesHealthCheckRequest(config ai.ProviderConfig) (*http.Request, error) {
+	body := map[string]interface{}{
+		"model":      config.Model,
+		"max_tokens": 1,
+		"messages": []map[string]string{
+			{"role": "user", "content": "ping"},
+		},
+	}
+	bodyBytes, err := json.Marshal(body)
+	if err != nil {
+		return nil, fmt.Errorf("序列化请求失败: %w", err)
+	}
+	req, err := http.NewRequest("POST", resolveAnthropicMessagesURL(config.BaseURL), strings.NewReader(string(bodyBytes)))
+	if err != nil {
+		return nil, fmt.Errorf("创建请求失败: %w", err)
+	}
+	req.Header.Set("Content-Type", "application/json")
+	provider.ApplyAnthropicAuthHeaders(req.Header, config.BaseURL, config.APIKey)
+	for k, v := range config.Headers {
+		req.Header.Set(k, v)
+	}
+	return req, nil
+}
+
 // AISetActiveProvider 设置活动 Provider
 func (s *Service) AISetActiveProvider(id string) {
 	s.mu.Lock()
@@ -380,7 +529,12 @@ func (s *Service) AIListModels() map[string]interface{} {
 		return map[string]interface{}{"success": false, "models": []string{}, "error": "未找到活跃 Provider"}
 	}

-	models, err := fetchModels(config)
+	config = normalizeProviderConfig(config)
+	if staticModels := defaultStaticModelsForProvider(config); len(staticModels) > 0 {
+		return map[string]interface{}{"success": true, "models": staticModels, "source": "static"}
+	}
+
+	models, err := fetchModelsFunc(config)
 	if err != nil {
 		// 回退到配置中的静态模型列表
 		if len(config.Models) > 0 {
@@ -389,10 +543,17 @@ func (s *Service) AIListModels() map[string]interface{} {
 		return map[string]interface{}{"success": false, "models": []string{}, "error": err.Error()}
 	}

+	models, err = filterFetchedModelsForProvider(config, models)
+	if err != nil {
+		return map[string]interface{}{"success": false, "models": []string{}, "error": err.Error()}
+	}
+
 	return map[string]interface{}{"success": true, "models": models, "source": "api"}
 }

 // fetchModels 从供应商 API 获取可用模型列表
+var fetchModelsFunc = fetchModels
+
 func fetchModels(config ai.ProviderConfig) ([]string, error) {
 	providerType := normalizedProviderType(config)
 	if staticModels := defaultStaticModelsForProvider(config); len(staticModels) > 0 {
@@ -588,8 +749,8 @@ func (s *Service) AIChatSend(messages []ai.Message, tools []ai.Tool) map[string]
 	}

 	return map[string]interface{}{
-		"success": true,
-		"content": resp.Content,
+		"success":    true,
+		"content":    resp.Content,
 		"tool_calls": resp.ToolCalls,
 		"tokensUsed": map[string]int{
 			"promptTokens":     resp.TokensUsed.PromptTokens,
--- a/internal/ai/service/service_qwen_test.go
+++ b/internal/ai/service/service_qwen_test.go
@@ -0,0 +1,157 @@
+package aiservice
+
+import (
+	"reflect"
+	"testing"
+
+	"GoNavi-Wails/internal/ai"
+)
+
+func TestDefaultStaticModelsForProvider_DoesNotReturnBailianStaticModels(t *testing.T) {
+	models := defaultStaticModelsForProvider(ai.ProviderConfig{
+		Type:    "anthropic",
+		BaseURL: "https://dashscope.aliyuncs.com/apps/anthropic",
+	})
+	if len(models) != 0 {
+		t.Fatalf("expected Bailian provider to rely on remote model list, got %v", models)
+	}
+}
+
+func TestDefaultStaticModelsForProvider_ReturnsDashScopeCodingPlanSupportedModels(t *testing.T) {
+	expected := []string{
+		"qwen3.5-plus",
+		"kimi-k2.5",
+		"glm-5",
+		"MiniMax-M2.5",
+		"qwen3-max-2026-01-23",
+		"qwen3-coder-next",
+		"qwen3-coder-plus",
+		"glm-4.7",
+	}
+	testCases := []ai.ProviderConfig{
+		{
+			Type:    "anthropic",
+			BaseURL: "https://coding.dashscope.aliyuncs.com/apps/anthropic",
+		},
+		{
+			Type:      "custom",
+			APIFormat: "claude-cli",
+			BaseURL:   "https://coding.dashscope.aliyuncs.com/apps/anthropic",
+		},
+	}
+
+	for _, testCase := range testCases {
+		models := defaultStaticModelsForProvider(testCase)
+		if !reflect.DeepEqual(models, expected) {
+			t.Fatalf("expected Coding Plan supported models %v, got %v for config %#v", expected, models, testCase)
+		}
+	}
+}
+
+func TestNormalizeProviderConfig_DoesNotForceModelForDashScopeProviders(t *testing.T) {
+	bailian := normalizeProviderConfig(ai.ProviderConfig{
+		Type:    "anthropic",
+		BaseURL: "https://dashscope.aliyuncs.com/apps/anthropic",
+	})
+	if bailian.Model != "" {
+		t.Fatalf("expected Bailian model to remain empty until explicit selection, got %q", bailian.Model)
+	}
+
+	codingPlan := normalizeProviderConfig(ai.ProviderConfig{
+		Type:    "anthropic",
+		BaseURL: "https://coding.dashscope.aliyuncs.com/apps/anthropic",
+	})
+	if codingPlan.Type != "custom" {
+		t.Fatalf("expected Coding Plan provider type to normalize to custom, got %q", codingPlan.Type)
+	}
+	if codingPlan.APIFormat != "claude-cli" {
+		t.Fatalf("expected Coding Plan provider api format to normalize to claude-cli, got %q", codingPlan.APIFormat)
+	}
+	if codingPlan.Model != "" {
+		t.Fatalf("expected Coding Plan model to remain empty until explicit selection, got %q", codingPlan.Model)
+	}
+	if len(codingPlan.Models) == 0 {
+		t.Fatal("expected Coding Plan provider to expose official supported models")
+	}
+	if codingPlan.Models[0] != "qwen3.5-plus" {
+		t.Fatalf("expected Coding Plan provider to expose latest supported models, got %v", codingPlan.Models)
+	}
+}
+
+func TestResolveModelsURL_UsesDashScopeCompatibleModelsEndpointForBailianAnthropic(t *testing.T) {
+	url := resolveModelsURL(ai.ProviderConfig{
+		Type:    "anthropic",
+		BaseURL: "https://dashscope.aliyuncs.com/apps/anthropic",
+	})
+	if url != "https://dashscope.aliyuncs.com/compatible-mode/v1/models" {
+		t.Fatalf("expected Bailian models endpoint, got %q", url)
+	}
+}
+
+func TestAIListModels_ReturnsStaticModelsForDashScopeCodingPlanWithoutRemoteFetch(t *testing.T) {
+	originalFetchModelsFunc := fetchModelsFunc
+	fetchModelsFunc = func(config ai.ProviderConfig) ([]string, error) {
+		t.Fatalf("expected Coding Plan model list to stay static and skip remote fetch, got config %#v", config)
+		return nil, nil
+	}
+	defer func() {
+		fetchModelsFunc = originalFetchModelsFunc
+	}()
+
+	service := NewService()
+	service.providers = []ai.ProviderConfig{
+		{
+			ID:      "provider-coding-plan",
+			Type:    "anthropic",
+			BaseURL: "https://coding.dashscope.aliyuncs.com/apps/anthropic",
+		},
+	}
+	service.activeProvider = "provider-coding-plan"
+
+	result := service.AIListModels()
+	if result["success"] != true {
+		t.Fatalf("expected AIListModels to succeed, got %#v", result)
+	}
+	models, ok := result["models"].([]string)
+	if !ok {
+		t.Fatalf("expected []string models, got %#v", result["models"])
+	}
+	if len(models) == 0 || models[0] != "qwen3.5-plus" {
+		t.Fatalf("expected official static Coding Plan models, got %#v", models)
+	}
+	if source, _ := result["source"].(string); source != "static" {
+		t.Fatalf("expected static source, got %#v", result["source"])
+	}
+}
+
+func TestAITestProvider_UsesClaudeCLIHealthCheckForDashScopeCodingPlan(t *testing.T) {
+	originalClaudeCLIHealthCheckFunc := claudeCLIHealthCheckFunc
+	defer func() {
+		claudeCLIHealthCheckFunc = originalClaudeCLIHealthCheckFunc
+	}()
+
+	var received ai.ProviderConfig
+	claudeCLIHealthCheckFunc = func(config ai.ProviderConfig) error {
+		received = config
+		return nil
+	}
+
+	service := NewService()
+	result := service.AITestProvider(ai.ProviderConfig{
+		Type:    "anthropic",
+		BaseURL: "https://coding.dashscope.aliyuncs.com/apps/anthropic",
+		APIKey:  "sk-test",
+	})
+	if result["success"] != true {
+		t.Fatalf("expected AITestProvider to succeed, got %#v", result)
+	}
+	if received.Type != "custom" {
+		t.Fatalf("expected Coding Plan test to use custom provider type, got %q", received.Type)
+	}
+	if received.APIFormat != "claude-cli" {
+		t.Fatalf("expected Coding Plan test to use claude-cli api format, got %q", received.APIFormat)
+	}
+	if received.Model != "qwen3.5-plus" {
+		t.Fatalf("expected Coding Plan test to default probe model to qwen3.5-plus, got %q", received.Model)
+	}
+}
--- a/internal/ai/service/service_test.go
+++ b/internal/ai/service/service_test.go
@@ -37,6 +37,43 @@ func TestResolveModelsURL_UsesOpenAIModelsEndpointForOpenAICompatibleProvider(t
 	}
 }

+func TestResolveModelsURL_UsesVersionedVolcengineCodingPlanPath(t *testing.T) {
+	url := resolveModelsURL(ai.ProviderConfig{
+		Type:    "openai",
+		BaseURL: "https://ark.cn-beijing.volces.com/api/coding/v3",
+	})
+	if url != "https://ark.cn-beijing.volces.com/api/coding/v3/models" {
+		t.Fatalf("expected volcengine coding plan models endpoint, got %q", url)
+	}
+}
+
+func TestResolveModelsURL_UsesVersionedZhipuPath(t *testing.T) {
+	url := resolveModelsURL(ai.ProviderConfig{
+		Type:    "openai",
+		BaseURL: "https://open.bigmodel.cn/api/paas/v4",
+	})
+	if url != "https://open.bigmodel.cn/api/paas/v4/models" {
+		t.Fatalf("expected zhipu models endpoint, got %q", url)
+	}
+}
+
+func TestNewModelsRequest_StripsChatCompletionsSuffixForOpenAICompatibleProvider(t *testing.T) {
+	req, err := newModelsRequest(ai.ProviderConfig{
+		Type:    "openai",
+		BaseURL: "https://ark.cn-beijing.volces.com/api/v3/chat/completions",
+		APIKey:  "sk-test",
+	})
+	if err != nil {
+		t.Fatalf("unexpected error: %v", err)
+	}
+	if req.URL.String() != "https://ark.cn-beijing.volces.com/api/v3/models" {
+		t.Fatalf("expected normalized models endpoint, got %q", req.URL.String())
+	}
+	if got := req.Header.Get("Authorization"); got != "Bearer sk-test" {
+		t.Fatalf("expected bearer auth header, got %q", got)
+	}
+}
+
 func TestDefaultStaticModelsForProvider_ReturnsMiniMaxAnthropicModels(t *testing.T) {
 	models := defaultStaticModelsForProvider(ai.ProviderConfig{
 		Type:    "anthropic",
@@ -56,6 +93,16 @@ func TestDefaultStaticModelsForProvider_ReturnsMiniMaxAnthropicModels(t *testing
 	}
 }

+func TestDefaultStaticModelsForProvider_DoesNotReturnDashScopeBailianStaticModels(t *testing.T) {
+	models := defaultStaticModelsForProvider(ai.ProviderConfig{
+		Type:    "anthropic",
+		BaseURL: "https://dashscope.aliyuncs.com/apps/anthropic",
+	})
+	if len(models) != 0 {
+		t.Fatalf("expected Bailian provider to fetch models remotely, got %v", models)
+	}
+}
+
 func TestNewProviderHealthCheckRequest_UsesMessagesEndpointForMiniMaxAnthropic(t *testing.T) {
 	req, err := newProviderHealthCheckRequest(ai.ProviderConfig{
 		Type:    "anthropic",
@@ -76,3 +123,30 @@ func TestNewProviderHealthCheckRequest_UsesMessagesEndpointForMiniMaxAnthropic(t
 		t.Fatalf("expected x-api-key header to be set, got %q", got)
 	}
 }
+
+func TestNewProviderHealthCheckRequest_UsesMessagesEndpointForDashScopeAnthropic(t *testing.T) {
+	req, err := newProviderHealthCheckRequest(ai.ProviderConfig{
+		Type:    "anthropic",
+		BaseURL: "https://dashscope.aliyuncs.com/apps/anthropic",
+		Model:   "qwen3.5-plus",
+		APIKey:  "sk-test",
+	})
+	if err != nil {
+		t.Fatalf("unexpected error: %v", err)
+	}
+	if req.Method != "POST" {
+		t.Fatalf("expected POST request, got %s", req.Method)
+	}
+	if req.URL.String() != "https://dashscope.aliyuncs.com/apps/anthropic/v1/messages" {
+		t.Fatalf("expected DashScope messages endpoint, got %q", req.URL.String())
+	}
+	if got := req.Header.Get("x-api-key"); got != "sk-test" {
+		t.Fatalf("expected x-api-key header to be set, got %q", got)
+	}
+	if got := req.Header.Get("Authorization"); got != "Bearer sk-test" {
+		t.Fatalf("expected bearer authorization header, got %q", got)
+	}
+	if got := req.Header.Get("anthropic-version"); got != "" {
+		t.Fatalf("expected no anthropic-version header for DashScope, got %q", got)
+	}
+}
--- a/internal/ai/service/service_volcengine_test.go
+++ b/internal/ai/service/service_volcengine_test.go
@@ -0,0 +1,111 @@
+package aiservice
+
+import (
+	"reflect"
+	"strings"
+	"testing"
+
+	"GoNavi-Wails/internal/ai"
+)
+
+func TestIsVolcengineCodingPlanProvider_MatchesCodingPlanBaseURL(t *testing.T) {
+	if !isVolcengineCodingPlanProvider(ai.ProviderConfig{
+		Type:    "openai",
+		BaseURL: "https://ark.cn-beijing.volces.com/api/coding/v3",
+	}) {
+		t.Fatal("expected volcengine coding plan provider to be detected")
+	}
+}
+
+func TestFilterVolcengineCodingPlanModels_KeepsOnlySupportedFamilies(t *testing.T) {
+	filtered := filterVolcengineCodingPlanModels([]string{
+		"Auto",
+		"qwen3-14b-20250429",
+		"wan2-1-14b-t2v-250225",
+		"Doubao-Seed-2.0-Code",
+		"Doubao-Seed-2.0-pro",
+		"Doubao-Seed-2.0-lite",
+		"doubao-seed-code-32k-250615",
+		"MiniMax-M2.5",
+		"GLM-4.7",
+		"DeepSeek-V3.2",
+		"kimi-k2-turbo-preview",
+	})
+
+	expected := []string{
+		"Auto",
+		"Doubao-Seed-2.0-Code",
+		"Doubao-Seed-2.0-pro",
+		"Doubao-Seed-2.0-lite",
+		"doubao-seed-code-32k-250615",
+		"MiniMax-M2.5",
+		"GLM-4.7",
+		"DeepSeek-V3.2",
+		"kimi-k2-turbo-preview",
+	}
+	if !reflect.DeepEqual(filtered, expected) {
+		t.Fatalf("expected filtered models %v, got %v", expected, filtered)
+	}
+}
+
+func TestFilterVolcengineCodingPlanModels_DoesNotBroadlyMatchAutoKeyword(t *testing.T) {
+	filtered := filterVolcengineCodingPlanModels([]string{
+		"Auto",
+		"automatic-router-preview",
+	})
+
+	expected := []string{"Auto"}
+	if !reflect.DeepEqual(filtered, expected) {
+		t.Fatalf("expected only exact Auto model to remain, got %v", filtered)
+	}
+}
+
+func TestFilterFetchedModelsForProvider_DoesNotFilterVolcengineArk(t *testing.T) {
+	rawModels := []string{
+		"qwen3-14b-20250429",
+		"wan2-1-14b-t2v-250225",
+	}
+
+	filtered, err := filterFetchedModelsForProvider(ai.ProviderConfig{
+		Type:    "openai",
+		BaseURL: "https://ark.cn-beijing.volces.com/api/v3",
+	}, rawModels)
+	if err != nil {
+		t.Fatalf("unexpected error: %v", err)
+	}
+	if !reflect.DeepEqual(filtered, rawModels) {
+		t.Fatalf("expected ark models to stay untouched, got %v", filtered)
+	}
+}
+
+func TestAIListModels_ReturnsFailureWhenVolcengineCodingPlanModelsAreFilteredEmpty(t *testing.T) {
+	originalFetchModelsFunc := fetchModelsFunc
+	fetchModelsFunc = func(config ai.ProviderConfig) ([]string, error) {
+		return []string{
+			"qwen3-14b-20250429",
+			"wan2-1-14b-t2v-250225",
+		}, nil
+	}
+	defer func() {
+		fetchModelsFunc = originalFetchModelsFunc
+	}()
+
+	service := NewService()
+	service.providers = []ai.ProviderConfig{
+		{
+			ID:      "provider-coding",
+			Type:    "openai",
+			BaseURL: "https://ark.cn-beijing.volces.com/api/coding/v3",
+		},
+	}
+	service.activeProvider = "provider-coding"
+
+	result := service.AIListModels()
+	if result["success"] != false {
+		t.Fatalf("expected AIListModels to fail, got %#v", result)
+	}
+	errorMessage, _ := result["error"].(string)
+	if !strings.Contains(errorMessage, "当前接口未返回可用的火山 Coding Plan 模型") {
+		t.Fatalf("expected specific coding plan error, got %q", errorMessage)
+	}
+}
--- a/internal/ai/types.go
+++ b/internal/ai/types.go
@@ -25,7 +25,7 @@ type Tool struct {

 // Message 表示一条对话消息
 type Message struct {
-	Role       string     `json:"role"`    // "system" | "user" | "assistant" | "tool"
+	Role       string     `json:"role"` // "system" | "user" | "assistant" | "tool"
 	Content    string     `json:"content"`
 	Images     []string   `json:"images,omitempty"`       // base64 encoded images with data:image/png;base64,... prefix
 	ToolCallID string     `json:"tool_call_id,omitempty"` // 当 role 为 "tool" 时必须传递
@@ -66,13 +66,13 @@ type StreamChunk struct {
 // ProviderConfig AI Provider 配置
 type ProviderConfig struct {
 	ID          string            `json:"id"`
-	Type        string            `json:"type"`        // openai | anthropic | gemini | custom
+	Type        string            `json:"type"` // openai | anthropic | gemini | custom
 	Name        string            `json:"name"`
 	APIKey      string            `json:"apiKey"`
 	BaseURL     string            `json:"baseUrl"`
 	Model       string            `json:"model"`
 	Models      []string          `json:"models,omitempty"`
-	APIFormat   string            `json:"apiFormat,omitempty"` // custom 专用: openai | anthropic | gemini
+	APIFormat   string            `json:"apiFormat,omitempty"` // custom 专用: openai | anthropic | gemini | claude-cli
 	Headers     map[string]string `json:"headers,omitempty"`
 	MaxTokens   int               `json:"maxTokens"`
 	Temperature float64           `json:"temperature"`
--- a/internal/app/app.go
+++ b/internal/app/app.go
@@ -64,9 +64,14 @@ func NewApp() *App {
 	}
 }

-// Startup is called when the app starts. The context is saved
-// so we can call the runtime methods
-func (a *App) Startup(ctx context.Context) {
+// InitializeLifecycle attaches runtime context without exposing lifecycle internals to Wails bindings.
+func InitializeLifecycle(a *App, ctx context.Context) {
+	a.startup(ctx)
+}
+
+// startup is called when the app starts. The context is saved
+// so we can call the runtime methods.
+func (a *App) startup(ctx context.Context) {
 	a.ctx = ctx
 	a.startedAt = time.Now()
 	logger.Init()
@@ -603,7 +608,7 @@ func (a *App) connectDatabaseWithStartupRetry(rawConfig connection.ConnectionCon

 		if err := dbInst.Connect(connectConfig); err == nil {
 			if attempt > 1 {
-				logger.Warnf("数据库连接在启动保护重试后成功：%s 缓存Key=%s 尝试=%d/%d", formatConnSummary(effectiveConfig), cacheKey, attempt, startupConnectRetryAttempts)
+				logger.Warnf("数据库连接在重试后成功：%s 缓存Key=%s 尝试=%d/%d", formatConnSummary(effectiveConfig), cacheKey, attempt, startupConnectRetryAttempts)
 			}
 			return dbInst, effectiveConfig, nil
 		} else {
@@ -611,10 +616,10 @@ func (a *App) connectDatabaseWithStartupRetry(rawConfig connection.ConnectionCon
 			wrapped := wrapConnectError(effectiveConfig, err)
 			lastErr = wrapped
 			logger.Error(wrapped, "建立数据库连接失败：%s 缓存Key=%s", formatConnSummary(effectiveConfig), cacheKey)
-			if !a.shouldRetryStartupConnect(err, attempt) {
+			if !a.shouldRetryConnect(err, attempt) {
 				return nil, effectiveConfig, wrapped
 			}
-			logger.Warnf("检测到启动期瞬时网络失败，准备重试连接：%s 缓存Key=%s 尝试=%d/%d 延迟=%s 原因=%s",
+			logger.Warnf("检测到瞬时网络失败，准备重试连接：%s 缓存Key=%s 尝试=%d/%d 延迟=%s 原因=%s",
 				formatConnSummary(effectiveConfig), cacheKey, attempt, startupConnectRetryAttempts, startupConnectRetryDelay, normalizeErrorMessage(err))
 			time.Sleep(startupConnectRetryDelay)
 		}
@@ -645,18 +650,21 @@ func (a *App) startupPhaseLabel() string {
 	return fmt.Sprintf("稳定期(age=%s)", age)
 }

-func (a *App) shouldRetryStartupConnect(err error, attempt int) bool {
+func (a *App) shouldRetryConnect(err error, attempt int) bool {
 	if attempt >= startupConnectRetryAttempts {
 		return false
 	}
-	if a == nil || a.startedAt.IsZero() {
+	if !isTransientStartupConnectError(err) {
 		return false
 	}
-	age := time.Since(a.startedAt)
-	if age < 0 || age > startupConnectRetryWindow {
-		return false
+	if a != nil && !a.startedAt.IsZero() {
+		age := time.Since(a.startedAt)
+		if age >= 0 && age <= startupConnectRetryWindow {
+			return true
+		}
 	}
-	return isTransientStartupConnectError(err)
+	// Outside startup window, still grant one retry for transient network glitches.
+	return attempt == 1
 }

 func isTransientStartupConnectError(err error) bool {
@@ -700,8 +708,8 @@ func (a *App) CancelQuery(queryID string) connection.QueryResult {
 	return connection.QueryResult{Success: false, Message: "查询不存在或已完成"}
 }

-// CleanupStaleQueries removes queries older than maxAge
-func (a *App) CleanupStaleQueries(maxAge time.Duration) {
+// cleanupStaleQueries removes queries older than maxAge.
+func (a *App) cleanupStaleQueries(maxAge time.Duration) {
 	a.queryMu.Lock()
 	defer a.queryMu.Unlock()

--- a/internal/app/app_startup_connect_retry_test.go
+++ b/internal/app/app_startup_connect_retry_test.go
@@ -2,11 +2,14 @@ package app

 import (
 	"errors"
+	"os"
+	"strings"
 	"testing"
 	"time"

 	"GoNavi-Wails/internal/connection"
 	"GoNavi-Wails/internal/db"
+	"GoNavi-Wails/internal/logger"
 )

 type fakeStartupRetryDB struct {
@@ -106,7 +109,7 @@ func TestConnectDatabaseWithStartupRetry_RetriesTransientFailureAndReappliesGlob
 	}
 }

-func TestConnectDatabaseWithStartupRetry_DoesNotRetryOutsideStartupWindow(t *testing.T) {
+func TestConnectDatabaseWithStartupRetry_RetriesOnceOutsideStartupWindow(t *testing.T) {
 	originalNewDatabaseFunc := newDatabaseFunc
 	originalResolveDialConfigWithProxyFunc := resolveDialConfigWithProxyFunc
 	defer func() {
@@ -130,12 +133,165 @@ func TestConnectDatabaseWithStartupRetry_DoesNotRetryOutsideStartupWindow(t *tes
 	a := &App{startedAt: time.Now().Add(-startupConnectRetryWindow - time.Second)}
 	rawConfig := connection.ConnectionConfig{Type: "postgres", Host: "10.1.131.86", Port: 5432, User: "postgres"}

+	_, _, err := a.connectDatabaseWithStartupRetry(rawConfig)
+	if err == nil {
+		t.Fatal("expected error, got nil")
+	}
+	if connectCalls != 2 {
+		t.Fatalf("expected 2 connect attempts outside startup window, got %d", connectCalls)
+	}
+}
+
+func TestConnectDatabaseWithStartupRetry_DoesNotRetryOutsideStartupWindowForNonTransientError(t *testing.T) {
+	originalNewDatabaseFunc := newDatabaseFunc
+	originalResolveDialConfigWithProxyFunc := resolveDialConfigWithProxyFunc
+	defer func() {
+		newDatabaseFunc = originalNewDatabaseFunc
+		resolveDialConfigWithProxyFunc = originalResolveDialConfigWithProxyFunc
+	}()
+
+	connectCalls := 0
+	newDatabaseFunc = func(dbType string) (db.Database, error) {
+		return &fakeStartupRetryDB{
+			connect: func(config connection.ConnectionConfig) error {
+				connectCalls++
+				return errors.New("pq: password authentication failed")
+			},
+		}, nil
+	}
+	resolveDialConfigWithProxyFunc = func(raw connection.ConnectionConfig) (connection.ConnectionConfig, error) {
+		return raw, nil
+	}
+
+	a := &App{startedAt: time.Now().Add(-startupConnectRetryWindow - time.Second)}
+	rawConfig := connection.ConnectionConfig{Type: "postgres", Host: "10.1.131.86", Port: 5432, User: "postgres"}
+
 	_, _, err := a.connectDatabaseWithStartupRetry(rawConfig)
 	if err == nil {
 		t.Fatal("expected error, got nil")
 	}
 	if connectCalls != 1 {
-		t.Fatalf("expected 1 connect attempt outside startup window, got %d", connectCalls)
+		t.Fatalf("expected 1 connect attempt outside startup window for non-transient error, got %d", connectCalls)
+	}
+}
+
+func TestConnectDatabaseWithStartupRetry_LogsRetryHintOutsideStartupWindow(t *testing.T) {
+	originalNewDatabaseFunc := newDatabaseFunc
+	originalResolveDialConfigWithProxyFunc := resolveDialConfigWithProxyFunc
+	defer func() {
+		newDatabaseFunc = originalNewDatabaseFunc
+		resolveDialConfigWithProxyFunc = originalResolveDialConfigWithProxyFunc
+	}()
+
+	logPath := logger.Path()
+	beforeSize := int64(0)
+	if fi, err := os.Stat(logPath); err == nil {
+		beforeSize = fi.Size()
+	}
+
+	connectCalls := 0
+	newDatabaseFunc = func(dbType string) (db.Database, error) {
+		return &fakeStartupRetryDB{
+			connect: func(config connection.ConnectionConfig) error {
+				connectCalls++
+				if connectCalls == 1 {
+					return errors.New("dial tcp 10.1.131.86:5432: connect: no route to host")
+				}
+				return nil
+			},
+		}, nil
+	}
+	resolveDialConfigWithProxyFunc = func(raw connection.ConnectionConfig) (connection.ConnectionConfig, error) {
+		return raw, nil
+	}
+
+	a := &App{startedAt: time.Now().Add(-startupConnectRetryWindow - time.Second)}
+	rawConfig := connection.ConnectionConfig{Type: "postgres", Host: "10.1.131.86", Port: 5432, User: "postgres"}
+
+	_, _, err := a.connectDatabaseWithStartupRetry(rawConfig)
+	if err != nil {
+		t.Fatalf("expected success after retry, got error: %v", err)
+	}
+	if connectCalls != 2 {
+		t.Fatalf("expected 2 connect attempts, got %d", connectCalls)
+	}
+
+	logContent, readErr := os.ReadFile(logPath)
+	if readErr != nil {
+		t.Fatalf("read log failed: %v", readErr)
+	}
+	if int64(len(logContent)) < beforeSize {
+		t.Fatalf("expected log file to grow, before=%d after=%d", beforeSize, len(logContent))
+	}
+	appended := string(logContent[beforeSize:])
+	if !strings.Contains(appended, "检测到瞬时网络失败，准备重试连接") {
+		t.Fatalf("expected retry hint log in appended segment, got: %s", appended)
+	}
+}
+
+func TestConnectDatabaseWithStartupRetry_OutsideStartupWindowTransientFailureStopsAfterOneRetry(t *testing.T) {
+	originalNewDatabaseFunc := newDatabaseFunc
+	originalResolveDialConfigWithProxyFunc := resolveDialConfigWithProxyFunc
+	defer func() {
+		newDatabaseFunc = originalNewDatabaseFunc
+		resolveDialConfigWithProxyFunc = originalResolveDialConfigWithProxyFunc
+	}()
+
+	connectCalls := 0
+	newDatabaseFunc = func(dbType string) (db.Database, error) {
+		return &fakeStartupRetryDB{
+			connect: func(config connection.ConnectionConfig) error {
+				connectCalls++
+				return errors.New("dial tcp 10.1.131.86:5432: connect: no route to host")
+			},
+		}, nil
+	}
+	resolveDialConfigWithProxyFunc = func(raw connection.ConnectionConfig) (connection.ConnectionConfig, error) {
+		return raw, nil
+	}
+
+	a := &App{startedAt: time.Now().Add(-startupConnectRetryWindow - time.Second)}
+	rawConfig := connection.ConnectionConfig{Type: "postgres", Host: "10.1.131.86", Port: 5432, User: "postgres"}
+
+	_, _, err := a.connectDatabaseWithStartupRetry(rawConfig)
+	if err == nil {
+		t.Fatal("expected error, got nil")
+	}
+	if connectCalls != 2 {
+		t.Fatalf("expected 2 connect attempts outside startup window for transient error, got %d", connectCalls)
+	}
+}
+
+func TestConnectDatabaseWithStartupRetry_StartupWindowTransientFailureUsesFullRetryBudget(t *testing.T) {
+	originalNewDatabaseFunc := newDatabaseFunc
+	originalResolveDialConfigWithProxyFunc := resolveDialConfigWithProxyFunc
+	defer func() {
+		newDatabaseFunc = originalNewDatabaseFunc
+		resolveDialConfigWithProxyFunc = originalResolveDialConfigWithProxyFunc
+	}()
+
+	connectCalls := 0
+	newDatabaseFunc = func(dbType string) (db.Database, error) {
+		return &fakeStartupRetryDB{
+			connect: func(config connection.ConnectionConfig) error {
+				connectCalls++
+				return errors.New("dial tcp 10.1.131.86:5432: connect: no route to host")
+			},
+		}, nil
+	}
+	resolveDialConfigWithProxyFunc = func(raw connection.ConnectionConfig) (connection.ConnectionConfig, error) {
+		return raw, nil
+	}
+
+	a := &App{startedAt: time.Now()}
+	rawConfig := connection.ConnectionConfig{Type: "postgres", Host: "10.1.131.86", Port: 5432, User: "postgres"}
+
+	_, _, err := a.connectDatabaseWithStartupRetry(rawConfig)
+	if err == nil {
+		t.Fatal("expected error, got nil")
+	}
+	if connectCalls != startupConnectRetryAttempts {
+		t.Fatalf("expected %d connect attempts in startup window, got %d", startupConnectRetryAttempts, connectCalls)
 	}
 }

--- a/internal/app/methods_db_cancel_test.go
+++ b/internal/app/methods_db_cancel_test.go
@@ -88,7 +88,7 @@ func TestCleanupStaleQueries(t *testing.T) {
 	app.queryMu.Unlock()

 	// Cleanup queries older than 1 hour
-	app.CleanupStaleQueries(1 * time.Hour)
+	app.cleanupStaleQueries(1 * time.Hour)

 	// Verify stale query was removed
 	app.queryMu.Lock()
@@ -110,7 +110,7 @@ func TestCleanupStaleQueries(t *testing.T) {
 	defer cancel2()

 	// Cleanup queries older than 1 hour
-	app.CleanupStaleQueries(1 * time.Hour)
+	app.cleanupStaleQueries(1 * time.Hour)

 	// Verify fresh query still exists
 	app.queryMu.Lock()
--- a/main.go
+++ b/main.go
@@ -34,8 +34,8 @@ func main() {
 		},
 		BackgroundColour: &options.RGBA{R: 0, G: 0, B: 0, A: 0},
 		OnStartup: func(ctx context.Context) {
-			application.Startup(ctx)
-			aiService.Startup(ctx)
+			app.InitializeLifecycle(application, ctx)
+			aiservice.InitializeLifecycle(aiService, ctx)
 		},
 		OnShutdown: application.Shutdown,
 		Bind: []interface{}{
Author	SHA1	Message	Date
Syngnat	09d013f27d	🐛 fix(app): 为稳定期首次连接增加瞬时网络重试保护 (#309 ) ## 问题背景在 app 启动后等待 20s 以上，再手动触发数据库连接时，遇到瞬时网络错误（如 `no route to host`）会立即失败，用户体感为“没有重试”。相关讨论与上下文参考： - https://github.com/Syngnat/GoNavi/pull/294 ## 问题描述此前重试保护逻辑主要以“应用启动窗口（20s）”为边界： - 启动窗口内：瞬时网络失败会自动重试 - 启动窗口外：即使是瞬时网络失败也不重试这导致“用户首次手动连接发生在稳定期”时，行为与预期不一致。 ## 本地复现关键日志（节选） ```log 2026/03/27 15:21:04.792462 [INFO] 应用启动完成（首次连接保护窗口=20s，最多重试=4 次） 2026/03/27 15:22:29.196794 [INFO] 获取数据库连接：... 启动阶段=稳定期(age=1m24.405s) 2026/03/27 15:22:29.208920 [ERROR] 建立数据库连接失败：... connect: no route to host 2026/03/27 15:22:29.212453 [ERROR] DBGetDatabases 获取连接失败：... connect: no route to host 2026/03/27 16:07:45.463959 [INFO] 获取数据库连接：... 启动阶段=稳定期(age=21s) 2026/03/27 16:07:45.470744 [ERROR] 建立数据库连接失败：... connect: no route to host 2026/03/27 16:07:45.473604 [WARN] 检测到瞬时网络失败，准备重试连接：... 尝试=1/4 延迟=800ms 2026/03/27 16:07:46.277658 [INFO] 获取数据库连接：... 启动阶段=稳定期(age=21.814s) 2026/03/27 16:07:46.281761 [INFO] 创建数据库驱动实例：... 尝试=2/4 2026/03/27 16:07:46.284741 [ERROR] 建立数据库连接失败：... connect: no route to host 2026/03/27 16:20:59.298636 [INFO] 应用启动完成（首次连接保护窗口=20s，最多重试=4 次） 2026/03/27 16:23:26.180978 [INFO] 获取数据库连接：... 启动阶段=稳定期(age=2m26.883s) 2026/03/27 16:23:26.201478 [INFO] 数据库连接成功并写入缓存：... ``` ## 变更内容 - 调整连接重试判定逻辑： - 启动窗口内：保持原有重试预算（最多 4 次） - 启动窗口外：若为瞬时网络错误，补充一次保护重试（总计 2 次尝试） - 非瞬时错误（如认证失败）在稳定期不重试 - 日志文案泛化，避免“仅启动期”误导： - 数据库连接在重试后成功 - 检测到瞬时网络失败，准备重试连接 ## 测试与验证 ### 新增/更新单元测试覆盖以下场景： - 启动期瞬时错误重试并成功 - 稳定期瞬时错误重试一次并成功 - 稳定期瞬时错误持续失败时，仅重试一次后停止 - 稳定期非瞬时错误不重试 - 稳定期重试路径输出重试提示日志 - 启动期瞬时错误失败时使用完整重试预算 ### 本地执行： - go test ./internal/app -run StartupRetry -count=1 - go test ./internal/app -count=1 ### 影响范围 - 连接建立重试策略（internal/app/app.go） - 启动重试相关测试（internal/app/app_startup_connect_retry_test.go） ## 风险与回滚 - 风险：稳定期瞬时网络错误会增加一次重试等待（约 800ms） - 回滚：可回退本 PR 即恢复“仅启动窗口重试”的旧策略	2026-03-27 17:30:14 +08:00
Syngnat	09aa526570	🐛 fix(ai/provider/chat-ui): 修复千问 Coding Plan 预设与 Claude CLI 报错 - 统一千问 Coding Plan 到 claude-cli 链路 - 修正旧配置识别与模型列表逻辑 - 透传 Claude CLI 鉴权失败和错误事件 - 移除误杀正常回复的启动定时器	2026-03-27 17:02:51 +08:00
DurianPankek	5844cd7c01	🐛 fix(app): 为稳定期首次连接增加瞬时网络重试保护	2026-03-27 16:27:46 +08:00
Syngnat	4f74c44147	🐛 fix(ai/provider/chat-ui): 修复AI供应商兼容性并优化聊天提示交互 - 修复通义千问百炼 Anthropic 兼容鉴权头与健康检查请求 - 拆分通义千问百炼通用与 Coding Plan 双入口，调整预设回填与模型策略 - 修复火山 Coding Plan 模型列表过滤逻辑，避免混入无关模型 - 统一 OpenAI 兼容供应商路径与模型列表处理，补充相关服务层测试 - 优化 AI 设置供应商卡片布局，统一高度并收紧文本展示 - 将聊天区模型校验提示改为输入框上方的内联提示卡，补充前端回归测试	2026-03-27 14:29:03 +08:00
Syngnat	a5fdfefa2d	🐛 fix(ai/volcengine): 修复火山引擎兼容路径并拆分双预设 - OpenAI 兼容 URL 归一化改为保留已有 v3 和 v4 版本段，避免火山与智谱地址被错误补 /v1 - 对误填 /chat/completions 和 /models 的地址先回退到 base URL，再拼接目标端点 - 模型列表与连通性检测复用统一端点解析逻辑，修复火山 Coding Plan 等兼容服务请求 - AI 设置页拆分火山方舟与火山 Coding Plan 两个预设，并按完整路径精确匹配回显 - 修正模型下拉默认值行为，未选模型时保持占位态，避免误用动态列表首项 - 补充 provider 与 service 回归测试，并新增需求追踪文档	2026-03-27 12:04:55 +08:00
Syngnat	37ac13b94e	🐛 fix(ai/wails-binding): 修复生命周期绑定生成类型错误 - 收敛 App 与 AI Service 的内部生命周期方法，避免被 Wails 误导出到前端 - 将启动初始化改为包级生命周期接线，保持主程序启动流程不变 - 隐藏内部清理方法，移除生成绑定中的无效 context/time 类型声明 - 同步更新 frontend/wailsjs 绑定文件，清理 Service 与 App 的错误导出 - 调整相关测试调用，确保内部方法重命名后行为一致	2026-03-27 11:42:57 +08:00