Merge branch 'feat-ai' into 'master'

refactor See merge request iesqa/httprunner!82
2026-05-07 05:32:43 +08:00 · 2025-05-07 12:03:17 +00:00
parent f6e421fc34 032a130418
commit ca0282a6bf
47 changed files with 1089 additions and 933 deletions
--- a/config.go
+++ b/config.go
@@ -4,7 +4,6 @@ import (
 	"reflect"

 	"github.com/httprunner/httprunner/v5/internal/builtin"
-	"github.com/httprunner/httprunner/v5/uixt/ai"
 	"github.com/httprunner/httprunner/v5/uixt/option"
 )

@@ -43,8 +42,8 @@ type TConfig struct {
 	Path              string                         `json:"path,omitempty" yaml:"path,omitempty"`     // testcase file path
 	PluginSetting     *PluginConfig                  `json:"plugin,omitempty" yaml:"plugin,omitempty"` // plugin config
 	IgnorePopup       bool                           `json:"ignore_popup,omitempty" yaml:"ignore_popup,omitempty"`
-	LLMService        ai.LLMServiceType              `json:"llm_service,omitempty" yaml:"llm_service,omitempty"`
-	CVService         ai.CVServiceType               `json:"cv_service,omitempty" yaml:"cv_service,omitempty"`
+	LLMService        option.LLMServiceType          `json:"llm_service,omitempty" yaml:"llm_service,omitempty"`
+	CVService         option.CVServiceType           `json:"cv_service,omitempty" yaml:"cv_service,omitempty"`
 }

 func (c *TConfig) Get() *TConfig {
@@ -112,13 +111,13 @@ func (c *TConfig) SetWeight(weight int) *TConfig {
 }

 // SetLLMService sets LLM service for current testcase.
-func (c *TConfig) SetLLMService(llmService ai.LLMServiceType) *TConfig {
+func (c *TConfig) SetLLMService(llmService option.LLMServiceType) *TConfig {
 	c.LLMService = llmService
 	return c
 }

 // SetCVService sets CV service for current testcase.
-func (c *TConfig) SetCVService(cvService ai.CVServiceType) *TConfig {
+func (c *TConfig) SetCVService(cvService option.CVServiceType) *TConfig {
 	c.CVService = cvService
 	return c
 }
--- a/convert/from_curl.go
+++ b/convert/from_curl.go
@@ -119,7 +119,7 @@ func LoadCurlCase(path string) (*hrp.TestCaseDef, error) {
 }

 func readFileLines(path string) ([]string, error) {
-	file, err := os.OpenFile(path, os.O_RDONLY, 0o600)
+	file, err := os.Open(path)
 	if err != nil {
 		log.Error().Err(err).Str("path", path).Msg("open file failed")
 		return nil, err
--- a/docs/uixt/ui_mark.md
+++ b/docs/uixt/ui_mark.md
@@ -0,0 +1,31 @@
+### UI操作标注
+
+针对 UI 操作的位置进行标注，帮助用户直观地了解操作发生的位置。
+
+#### 功能说明
+
+- 点击操作（tap）：使用红色矩形框标注点击位置
+- 滑动操作（swipe）：使用红色箭头标注滑动方向，从起始点指向结束点
+
+#### 使用方法
+
+只需在操作函数中添加 `WithMarkOperationEnabled(true)` 选项即可启用操作标注功能：
+
+```go
+// 启用操作标注功能
+opts := []option.ActionOption{option.WithMarkOperationEnabled(true)}
+
+// 执行点击操作，会自动用红色矩形标注点击位置
+err := driver.TapXY(0.5, 0.5, opts...)
+
+// 执行滑动操作，会自动用红色箭头标注滑动方向
+err = driver.Swipe(0.2, 0.5, 0.8, 0.5, opts...)
+
+// 可以同时使用其他选项
+opts = append(opts, option.WithScreenShotFileName("custom_name"))
+err = driver.TapXY(0.3, 0.7, opts...)
+```
+
+#### 标注结果
+
+标注后的图片会保存在截图目录中，文件名格式为：`{timestamp}_{tap|swipe}_marked.png`
--- a/examples/uitest/bili/android/cli.go
+++ b/examples/uitest/bili/android/cli.go
@@ -52,7 +52,7 @@ func launchAppDriver(pkgName string) (driverExt *uixt.XTDriver, err error) {

 	time.Sleep(15 * time.Second)

-	driverExt = uixt.NewXTDriver(driver)
+	driverExt, _ = uixt.NewXTDriver(driver)

 	// 处理弹窗
 	err = driverExt.ClosePopupsHandler()
--- a/examples/uitest/bili/ios/cli.go
+++ b/examples/uitest/bili/ios/cli.go
@@ -7,7 +7,6 @@ import (
 	"time"

 	"github.com/httprunner/httprunner/v5/uixt"
-	"github.com/httprunner/httprunner/v5/uixt/ai"
 	"github.com/httprunner/httprunner/v5/uixt/option"
 )

@@ -53,8 +52,8 @@ func launchAppDriver(pkgName string) (driverExt *uixt.XTDriver, err error) {

 	time.Sleep(15 * time.Second)

-	driverExt = uixt.NewXTDriver(driver,
-		ai.WithCVService(ai.CVServiceTypeVEDEM))
+	driverExt, _ = uixt.NewXTDriver(driver,
+		option.WithCVService(option.CVServiceTypeVEDEM))

 	// 处理弹窗
 	err = driverExt.ClosePopupsHandler()
--- a/examples/worldcup/cli.go
+++ b/examples/worldcup/cli.go
@@ -8,7 +8,6 @@ import (
 	"github.com/spf13/cobra"

 	"github.com/httprunner/httprunner/v5/uixt"
-	"github.com/httprunner/httprunner/v5/uixt/ai"
 	"github.com/httprunner/httprunner/v5/uixt/option"
 )

@@ -39,8 +38,8 @@ var rootCmd = &cobra.Command{
 			return errors.New("android or ios app bundldID is required")
 		}

-		driverExt := uixt.NewXTDriver(driver,
-			ai.WithCVService(ai.CVServiceTypeVEDEM),
+		driverExt, _ := uixt.NewXTDriver(driver,
+			option.WithCVService(option.CVServiceTypeVEDEM),
 		)

 		wc := NewWorldCupLive(driverExt, matchName, bundleID, duration, interval)
--- a/examples/worldcup/main.go
+++ b/examples/worldcup/main.go
@@ -48,7 +48,7 @@ func initIOSDriver(uuid string) *uixt.XTDriver {
 		log.Fatal().Err(err).Msg("failed to init ios device")
 	}
 	driver, _ := device.NewDriver()
-	driverExt := uixt.NewXTDriver(driver)
+	driverExt, _ := uixt.NewXTDriver(driver)
 	return driverExt
 }

@@ -58,7 +58,7 @@ func initAndroidDriver(uuid string) *uixt.XTDriver {
 		log.Fatal().Err(err).Msg("failed to init android device")
 	}
 	driver, _ := device.NewDriver()
-	driverExt := uixt.NewXTDriver(driver)
+	driverExt, _ := uixt.NewXTDriver(driver)
 	return driverExt
 }

--- a/go.mod
+++ b/go.mod
@@ -9,7 +9,6 @@ require (
 	github.com/andybalholm/brotli v1.0.4
 	github.com/bytedance/sonic v1.13.2
 	github.com/cloudwego/eino v0.3.26
-	github.com/cloudwego/eino-ext/components/model/ark v0.1.6
 	github.com/cloudwego/eino-ext/components/model/openai v0.0.0-20250417123744-154d7ca4d3cd
 	github.com/cloudwego/eino-ext/components/tool/mcp v0.0.0-20250328102648-b47e7f1587fa
 	github.com/cloudwego/eino-ext/libs/acl/openai v0.0.0-20250417123744-154d7ca4d3cd
@@ -42,6 +41,7 @@ require (
 )

 require (
+	github.com/bytedance/mockey v1.2.14 // indirect
 	github.com/bytedance/sonic/loader v0.2.4 // indirect
 	github.com/cenkalti/backoff v2.2.1+incompatible // indirect
 	github.com/cloudwego/base64x v0.1.5 // indirect
@@ -100,8 +100,6 @@ require (
 	github.com/tadglines/go-pkgs v0.0.0-20210623144937-b983b20f54f9 // indirect
 	github.com/twitchyliquid64/golang-asm v0.15.1 // indirect
 	github.com/ugorji/go/codec v1.2.12 // indirect
-	github.com/volcengine/volc-sdk-golang v1.0.23 // indirect
-	github.com/volcengine/volcengine-go-sdk v1.0.185 // indirect
 	github.com/yargevad/filepathx v1.0.0 // indirect
 	github.com/yosida95/uritemplate/v3 v3.0.2 // indirect
 	go.mozilla.org/pkcs7 v0.0.0-20210826202110-33d05740a352 // indirect
--- a/go.sum
+++ b/go.sum
@@ -1,11 +1,8 @@
-cloud.google.com/go v0.26.0/go.mod h1:aQUYkXzVsufM+DwF1aE+0xfcU+56JwCaLick0ClmMTw=
-github.com/BurntSushi/toml v0.3.1/go.mod h1:xHWCNGjB5oqiDr8zfno3MHue2Ht5sIBksp03qcyfWMU=
 github.com/Masterminds/semver v1.5.0 h1:H65muMkzWKEuNDnfl9d70GUjFniHKHRbFPGBuZ3QEww=
 github.com/Masterminds/semver v1.5.0/go.mod h1:MB6lktGJrhw8PrUyiEoblNEGEQ+RzHPF078ddwwvV3Y=
 github.com/airbrake/gobrake v3.6.1+incompatible/go.mod h1:wM4gu3Cn0W0K7GUuVWnlXZU11AGBXMILnrdOU8Kn00o=
 github.com/andybalholm/brotli v1.0.4 h1:V7DdXeJtZscaqfNuAdSRuRFzuiKlHSC/Zh3zl9qY3JY=
 github.com/andybalholm/brotli v1.0.4/go.mod h1:fO7iG3H7G2nSZ7m0zPUDn85XEX2GTukHGRSepvi9Eig=
-github.com/avast/retry-go v3.0.0+incompatible/go.mod h1:XtSnn+n/sHqQIpZ10K1qAevBhOOCWBLXXy3hyiqqBrY=
 github.com/bitly/go-simplejson v0.5.0/go.mod h1:cXHtHw4XUPsvGaxgjIAn8PhEWG9NfngEKAMDJEczWVA=
 github.com/bmizerany/assert v0.0.0-20160611221934-b7ed37b82869/go.mod h1:Ekp36dRnpXw/yCqJaO+ZrUyxD+3VXMFFr56k5XYrpB4=
 github.com/bugsnag/bugsnag-go v1.4.0/go.mod h1:2oa8nejYd4cQ/b0hMIopN0lCRxU0bueqREvZLWFrtK8=
@@ -19,18 +16,14 @@ github.com/bytedance/sonic/loader v0.2.4 h1:ZWCw4stuXUsn1/+zQDqeE7JKP+QO47tz7QCN
 github.com/bytedance/sonic/loader v0.2.4/go.mod h1:N8A3vUdtUebEY2/VQC0MyhYeKUFosQU6FxH2JmUe6VI=
 github.com/cenkalti/backoff v2.2.1+incompatible h1:tNowT99t7UNflLxfYYSlKYsBpXdEet03Pg2g16Swow4=
 github.com/cenkalti/backoff v2.2.1+incompatible/go.mod h1:90ReRw6GdpyfrHakVjL/QHaoyV4aDUVVkXQJJJ3NXXM=
-github.com/census-instrumentation/opencensus-proto v0.2.1/go.mod h1:f6KPmirojxKA12rnyqOA5BBL4O983OfeGPqjHWSTneU=
 github.com/certifi/gocertifi v0.0.0-20190105021004-abcd57078448/go.mod h1:GJKEexRPVJrBSOjoqN5VNOIKJ5Q3RViH6eu3puDRwx4=
 github.com/chzyer/logex v1.1.10/go.mod h1:+Ywpsq7O8HXn0nuIou7OrIPyXbp3wmkHB+jjWRnGsAI=
 github.com/chzyer/readline v0.0.0-20180603132655-2972be24d48e/go.mod h1:nSuG5e5PlCu98SY8svDHJxuZscDgtXS6KTTbou5AhLI=
 github.com/chzyer/test v0.0.0-20180213035817-a1ea475d72b1/go.mod h1:Q3SI9o4m/ZMnBNeIyt5eFwwo7qiLfzFZmjNmxjkiQlU=
-github.com/client9/misspell v0.3.4/go.mod h1:qj6jICC3Q7zFZvVWo7KLAzC3yx5G7kyvSDkc90ppPyw=
 github.com/cloudwego/base64x v0.1.5 h1:XPciSp1xaq2VCSt6lF0phncD4koWyULpl5bUxbfCyP4=
 github.com/cloudwego/base64x v0.1.5/go.mod h1:0zlkT4Wn5C6NdauXdJRhSKRlJvmclQ1hhJgA0rcu/8w=
 github.com/cloudwego/eino v0.3.26 h1:FdJJTCdNrc9xPcYkLZiEyr7AA+WgyCKCbY+VNDXIaCE=
 github.com/cloudwego/eino v0.3.26/go.mod h1:wUjz990apdsaOraOXdh6CdhVXq8DJsOvLsVlxNTcNfY=
-github.com/cloudwego/eino-ext/components/model/ark v0.1.6 h1:k17Z9VIRBL0/t7Ty1drGgY9tVOraM5xuO6gy7Qx7xus=
-github.com/cloudwego/eino-ext/components/model/ark v0.1.6/go.mod h1:13kQjYGLMgla6xTbejlpqhuk3i5BPlNv5S+1pmknlOo=
 github.com/cloudwego/eino-ext/components/model/openai v0.0.0-20250417123744-154d7ca4d3cd h1:XEI7RezzV/cnOnhc1YeBJi6a0UoM41JTph4AZZR7+D8=
 github.com/cloudwego/eino-ext/components/model/openai v0.0.0-20250417123744-154d7ca4d3cd/go.mod h1:8gMakAGQUR+IaWTSD0cpcD4U5FYq5puZ73/QjXqs1oU=
 github.com/cloudwego/eino-ext/components/tool/mcp v0.0.0-20250328102648-b47e7f1587fa h1:Jrmw8Q9g1WcE+x5t3o0TsEBM8RoMRURJI6P52I/ld74=
@@ -53,8 +46,6 @@ github.com/dustin/go-humanize v1.0.1 h1:GzkhY7T5VNhEkwH0PVJgjz+fX1rhBrR7pRT3mDkp
 github.com/dustin/go-humanize v1.0.1/go.mod h1:Mu1zIs6XwVuF/gI1OepvI0qD18qycQx+mFykh5fBlto=
 github.com/elazarl/goproxy v0.0.0-20240726154733-8b0c20506380 h1:1NyRx2f4W4WBRyg0Kys0ZbaNmDDzZ2R/C7DTi+bbsJ0=
 github.com/elazarl/goproxy v0.0.0-20240726154733-8b0c20506380/go.mod h1:thX175TtLTzLj3p7N/Q9IiKZ7NF+p72cvL91emV0hzo=
-github.com/envoyproxy/go-control-plane v0.9.1-0.20191026205805-5f8ba28d4473/go.mod h1:YTl/9mNaCwkRvm6d1a2C3ymFceY/DCBVvsKhRF0iEA4=
-github.com/envoyproxy/protoc-gen-validate v0.1.0/go.mod h1:iSmxcyjqTsJpI2R4NaDN7+kN2VEUnK/pcBlmesArF7c=
 github.com/fatih/color v1.13.0/go.mod h1:kLAiJbzzSOZDVNGyDpeOxJ47H46qBXwg5ILebYFFOfk=
 github.com/fatih/color v1.16.0 h1:zmkK9Ngbjj+K0yRhTVONQh1p/HknKYSlNT+vZCzyokM=
 github.com/fatih/color v1.16.0/go.mod h1:fL2Sau1YI5c0pdGEVCbKQbLXB6edEj1ZgiY4NijnWvE=
@@ -106,27 +97,12 @@ github.com/goccy/go-json v0.10.2 h1:CrxCmQqYDkv1z7lO7Wbh2HN93uovUHgrECaO5ZrCXAU=
 github.com/goccy/go-json v0.10.2/go.mod h1:6MelG93GURQebXPDq3khkgXZkazVtN9CRI+MGFi0w8I=
 github.com/godbus/dbus/v5 v5.0.4/go.mod h1:xhWf0FNVPg57R7Z0UbKHbJfkEywrmjJnf7w5xrFpKfA=
 github.com/gofrs/uuid v3.2.0+incompatible/go.mod h1:b2aQJv3Z4Fp6yNu3cdSllBxTCLRxnplIgP/c0N/04lM=
-github.com/golang/glog v0.0.0-20160126235308-23def4e6c14b/go.mod h1:SBH7ygxi8pfUlaOkMMuAQtPIUF8ecWP5IEl/CR7VP2Q=
-github.com/golang/mock v1.1.1/go.mod h1:oTYuIxOrZwtPieC+H1uAHpcLFnEyAGVDL/k47Jfbm0A=
 github.com/golang/protobuf v1.2.0/go.mod h1:6lQm79b+lXiMfvg/cZm0SGofjICqVBUtrP5yJMmIC1U=
-github.com/golang/protobuf v1.3.2/go.mod h1:6lQm79b+lXiMfvg/cZm0SGofjICqVBUtrP5yJMmIC1U=
-github.com/golang/protobuf v1.4.0-rc.1/go.mod h1:ceaxUfeHdC40wWswd/P6IGgMaK3YpKi5j83Wpe3EHw8=
-github.com/golang/protobuf v1.4.0-rc.1.0.20200221234624-67d41d38c208/go.mod h1:xKAWHe0F5eneWXFV3EuXVDTCmh+JuBKY0li0aMyXATA=
-github.com/golang/protobuf v1.4.0-rc.2/go.mod h1:LlEzMj4AhA7rCAGe4KMBDvJI+AwstrUpVNzEA03Pprs=
-github.com/golang/protobuf v1.4.0-rc.4.0.20200313231945-b860323f09d0/go.mod h1:WU3c8KckQ9AFe+yFwt9sWVRKCVIyN9cPHBJSNnbL67w=
-github.com/golang/protobuf v1.4.0/go.mod h1:jodUvKwWbYaEsadDk5Fwe5c77LiNKVO9IDvqG2KuDX0=
-github.com/golang/protobuf v1.4.1/go.mod h1:U8fpvMrcmy5pZrNK1lt4xCsGvpyWQ/VVv6QDs8UjoX8=
-github.com/golang/protobuf v1.4.3/go.mod h1:oDoupMAO8OvCJWAcko0GGGIgR6R6ocIYbsSw735rRwI=
 github.com/golang/protobuf v1.5.0/go.mod h1:FsONVRAS9T7sI+LIUmWTfcYkHO4aIWwzhcaSAoJOfIk=
 github.com/golang/protobuf v1.5.3 h1:KhyjKVUg7Usr/dYsdSqoFveMYd5ko72D+zANwlG1mmg=
 github.com/golang/protobuf v1.5.3/go.mod h1:XVQd3VNwM+JqD3oG2Ue2ip4fOMUkwXdXDdiuN0vRsmY=
 github.com/google/btree v1.1.2 h1:xf4v41cLI2Z6FxbKm+8Bu+m8ifhj15JuZ9sa0jZCMUU=
 github.com/google/btree v1.1.2/go.mod h1:qOPhT0dTNdNzV6Z/lhRX0YXUafgPLFUh+gZMl761Gm4=
-github.com/google/go-cmp v0.2.0/go.mod h1:oXzfMopK8JAjlY9xF4vHSVASa0yLyX7SntLO5aqRK0M=
-github.com/google/go-cmp v0.3.0/go.mod h1:8QqcDgzrUqlUb/G2PQTWiueGozuR1884gddMywk6iLU=
-github.com/google/go-cmp v0.3.1/go.mod h1:8QqcDgzrUqlUb/G2PQTWiueGozuR1884gddMywk6iLU=
-github.com/google/go-cmp v0.4.0/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
-github.com/google/go-cmp v0.5.0/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
 github.com/google/go-cmp v0.5.5/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
 github.com/google/go-cmp v0.6.0 h1:ofyhxvXcZhMsU5ulbFiLKl/XBFqE1GSq7atu8tAmTRI=
 github.com/google/go-cmp v0.6.0/go.mod h1:17dUlkBOakJ0+DkrSSNjCkIjxS6bF9zb3elmeNGIjoY=
@@ -135,7 +111,6 @@ github.com/google/pprof v0.0.0-20210407192527-94a9f03dee38 h1:yAJXTCF9TqKcTiHJAE
 github.com/google/pprof v0.0.0-20210407192527-94a9f03dee38/go.mod h1:kpwsk12EmLew5upagYY7GY0pfYCcupk39gWOCRROcvE=
 github.com/google/shlex v0.0.0-20191202100458-e7afc7fbc510 h1:El6M4kTTCOh6aBiKaUGG7oYTSPP8MxqL4YI3kZKwcP4=
 github.com/google/shlex v0.0.0-20191202100458-e7afc7fbc510/go.mod h1:pupxD2MaaD3pAXIBCelhxNneeOaAeabZDe5s4K6zSpQ=
-github.com/google/uuid v1.3.0/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
 github.com/google/uuid v1.6.0 h1:NIvaJDMOsjHA8n1jAhLSgzrAzy1Hgr+hNrb57e+94F0=
 github.com/google/uuid v1.6.0/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
 github.com/goph/emperror v0.17.2 h1:yLapQcmEsO0ipe9p5TaN22djm3OFV/TfM/fcYP0/J18=
@@ -186,7 +161,6 @@ github.com/klauspost/cpuid/v2 v2.2.10/go.mod h1:hqwkgyIinND0mEev00jJYCxPNVRVXFQe
 github.com/knz/go-libedit v1.10.1/go.mod h1:MZTVkCWyz0oBc7JOWP3wNAzd002ZbM/5hgShxwh4x8M=
 github.com/konsorten/go-windows-terminal-sequences v1.0.1/go.mod h1:T0+1ngSBFLxvqU3pZ+m/2kptfBszLMUkC4ZK/EgS/cQ=
 github.com/kr/pretty v0.1.0/go.mod h1:dAy3ld7l9f0ibDNOQOHHMYYIIbhfbHSm3C4ZsoJORNo=
-github.com/kr/pretty v0.2.0/go.mod h1:ipq/a2n7PKx3OHsz4KJII5eveXtPO4qwEXGdVfWzfnI=
 github.com/kr/pretty v0.3.1 h1:flRD4NNwYAUpkphVc1HcthR4KEIFJ65n8Mw5qdRn3LE=
 github.com/kr/pretty v0.3.1/go.mod h1:hoEshYVHaxMs3cyo3Yncou5ZscifuDolrwPKZanG3xk=
 github.com/kr/pty v1.1.1/go.mod h1:pFQYn66WHrOpPYNljwOMqo10TkYh1fy3cYio2l3bCsQ=
@@ -259,7 +233,6 @@ github.com/pkg/errors v0.9.1 h1:FEBLx1zS214owpjy7qsBeixbURkuhQAwrK5UwLGTwt4=
 github.com/pkg/errors v0.9.1/go.mod h1:bwawxfHBFNV+L2hUp1rHADufV3IMtnDRdf1r5NINEl0=
 github.com/pmezard/go-difflib v1.0.0 h1:4DBwDE0NGyQoBHbLQYPwSUPoCMWR5BEzIk/f1lZbAQM=
 github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
-github.com/prometheus/client_model v0.0.0-20190812154241-14fe0d1b01d4/go.mod h1:xMI15A0UPsDsEKsMN9yxemIoYk6Tm2C1GtYGdfGttqA=
 github.com/quic-go/qtls-go1-20 v0.4.1 h1:D33340mCNDAIKBqXuAvexTNMUByrYmFYVfKfDN5nfFs=
 github.com/quic-go/qtls-go1-20 v0.4.1/go.mod h1:X9Nh97ZL80Z+bX/gUXMbipO6OxdiDi58b/fMC9mAL+k=
 github.com/quic-go/quic-go v0.40.1-0.20231203135336-87ef8ec48d55 h1:I4N3ZRnkZPbDN935Tg8QDf8fRpHp3bZ0U0/L42jBgNE=
@@ -299,7 +272,6 @@ github.com/stretchr/objx v0.5.2 h1:xuMeJ0Sdp5ZMRXx/aWO6RZxdr3beISkG5/G/aIRr3pY=
 github.com/stretchr/objx v0.5.2/go.mod h1:FRsXN1f5AsAjCGJKqEizvkpNtU+EGNCLh3NxZ/8L+MA=
 github.com/stretchr/testify v1.2.2/go.mod h1:a8OnRcib4nhh0OaRAV+Yts87kKdq0PP7pXfy6kDkUVs=
 github.com/stretchr/testify v1.3.0/go.mod h1:M5WIy9Dh21IEIfnGCwXGc5bZfKNJtfHm1UVUgZn+9EI=
-github.com/stretchr/testify v1.5.1/go.mod h1:5W2xD1RspED5o8YsWQXVCued0rvSQ+mT+I5cxcmMvtA=
 github.com/stretchr/testify v1.6.1/go.mod h1:6Fq8oRcR53rry900zMqJjRRixrwX3KX962/h/Wwjteg=
 github.com/stretchr/testify v1.7.0/go.mod h1:6Fq8oRcR53rry900zMqJjRRixrwX3KX962/h/Wwjteg=
 github.com/stretchr/testify v1.7.1/go.mod h1:6Fq8oRcR53rry900zMqJjRRixrwX3KX962/h/Wwjteg=
@@ -316,10 +288,6 @@ github.com/ugorji/go v1.2.7/go.mod h1:nF9osbDWLy6bDVv/Rtoh6QgnvNDpmCalQV5urGCCS6
 github.com/ugorji/go/codec v1.2.7/go.mod h1:WGN1fab3R1fzQlVQTkfxVtIBhWDRqOviHU95kRgeqEY=
 github.com/ugorji/go/codec v1.2.12 h1:9LC83zGrHhuUA9l16C9AHXAqEV/2wBQ4nkvumAE65EE=
 github.com/ugorji/go/codec v1.2.12/go.mod h1:UNopzCgEMSXjBc6AOMqYvWC1ktqTAfzJZUZgYf6w6lg=
-github.com/volcengine/volc-sdk-golang v1.0.23 h1:anOslb2Qp6ywnsbyq9jqR0ljuO63kg9PY+4OehIk5R8=
-github.com/volcengine/volc-sdk-golang v1.0.23/go.mod h1:AfG/PZRUkHJ9inETvbjNifTDgut25Wbkm2QoYBTbvyU=
-github.com/volcengine/volcengine-go-sdk v1.0.185 h1:MIH+YgdWZhO1fNg/vxLohl8ad7hlklaf46wpaTS1TN0=
-github.com/volcengine/volcengine-go-sdk v1.0.185/go.mod h1:gfEDc1s7SYaGoY+WH2dRrS3qiuDJMkwqyfXWCa7+7oA=
 github.com/x-cray/logrus-prefixed-formatter v0.5.2 h1:00txxvfBM9muc0jiLIEAkAcIMJzfthRT6usrui8uGmg=
 github.com/x-cray/logrus-prefixed-formatter v0.5.2/go.mod h1:2duySbKsL6M18s5GU7VPsoEPHyzalCE06qoARUCeBBE=
 github.com/yargevad/filepathx v1.0.0 h1:SYcT+N3tYGi+NvazubCNlvgIPbzAk7i7y2dwg3I5FYc=
@@ -338,20 +306,12 @@ golang.org/x/crypto v0.0.0-20191011191535-87dc89f01550/go.mod h1:yigFU9vqHzYiE8U
 golang.org/x/crypto v0.0.0-20220331220935-ae2d96664a29/go.mod h1:IxCIyHEi3zRg3s0A5j5BB6A9Jmi73HwBIUl50j+osU4=
 golang.org/x/crypto v0.37.0 h1:kJNSjF/Xp7kU0iB2Z+9viTPMW4EqqsrywMXLJOOsXSE=
 golang.org/x/crypto v0.37.0/go.mod h1:vg+k43peMZ0pUMhYmVAWysMK35e6ioLh3wB8ZCAfbVc=
-golang.org/x/exp v0.0.0-20190121172915-509febef88a4/go.mod h1:CJ0aWSM057203Lf6IL+f9T1iT9GByDxfZKAQTCR3kQA=
 golang.org/x/exp v0.0.0-20250408133849-7e4ce0ab07d0 h1:R84qjqJb5nVJMxqWYb3np9L5ZsaDtB+a39EqjV0JSUM=
 golang.org/x/exp v0.0.0-20250408133849-7e4ce0ab07d0/go.mod h1:S9Xr4PYopiDyqSyp5NjCrhFrqg6A5zA2E/iPHPhqnS8=
-golang.org/x/lint v0.0.0-20181026193005-c67002cb31c3/go.mod h1:UVdnD1Gm6xHRNCYTkRU2/jEulfH38KcIWyp/GAMgvoE=
-golang.org/x/lint v0.0.0-20190227174305-5b3e6a55c961/go.mod h1:wehouNa3lNwaWXcvxsM5YxQ5yQlVC4a0KAMCusXpPoU=
-golang.org/x/lint v0.0.0-20190313153728-d0100b6bd8b3/go.mod h1:6SW0HCj/g11FgYtHlgUYUwCkIfeOF89ocIRzGO/8vkc=
 golang.org/x/mod v0.1.1-0.20191105210325-c90efee705ee/go.mod h1:QqPTAvyqsEbceGzBzNggFXnrqF1CaUcvgkdR5Ot7KZg=
 golang.org/x/mod v0.24.0 h1:ZfthKaKaT4NrhGVZHO1/WDTwGES4De8KtWO0SIbNJMU=
 golang.org/x/mod v0.24.0/go.mod h1:IXM97Txy2VM4PJ3gI61r1YEk/gAj6zAHN3AdZt6S9Ww=
-golang.org/x/net v0.0.0-20180724234803-3673e40ba225/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
-golang.org/x/net v0.0.0-20180826012351-8a410e7b638d/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
 golang.org/x/net v0.0.0-20180906233101-161cd47e91fd/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
-golang.org/x/net v0.0.0-20190213061140-3a22650c66bd/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
-golang.org/x/net v0.0.0-20190311183353-d8887717615a/go.mod h1:t9HGtf8HONx5eT2rtn7q6eTqICYqUVnKs3thJo3Qplg=
 golang.org/x/net v0.0.0-20190404232315-eb5bcb51f2a3/go.mod h1:t9HGtf8HONx5eT2rtn7q6eTqICYqUVnKs3thJo3Qplg=
 golang.org/x/net v0.0.0-20190620200207-3b0461eec859/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
 golang.org/x/net v0.0.0-20190923162816-aa69164e4478/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
@@ -359,13 +319,10 @@ golang.org/x/net v0.0.0-20200114155413-6afb5195e5aa/go.mod h1:z5CRVTTTmAJ677TzLL
 golang.org/x/net v0.0.0-20211112202133-69e39bad7dc2/go.mod h1:9nx3DQGgdP8bBQD5qxJ1jj9UTztislL4KSBs9R2vV5Y=
 golang.org/x/net v0.39.0 h1:ZCu7HMWDxpXpaiKdhzIfaltL9Lp31x/3fCP11bc6/fY=
 golang.org/x/net v0.39.0/go.mod h1:X7NRbYVEA+ewNkCNyJ513WmMdQ3BineSwVtN2zD/d+E=
-golang.org/x/oauth2 v0.0.0-20180821212333-d2e6202438be/go.mod h1:N/0e6XlmueqKjAGxoOufVs8QHGRruUQn6yWY3a++T0U=
 golang.org/x/sync v0.0.0-20180314180146-1d60e4601c6f/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
-golang.org/x/sync v0.0.0-20181108010431-42b317875d0f/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
 golang.org/x/sync v0.0.0-20190423024810-112230192c58/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
 golang.org/x/sync v0.13.0 h1:AauUjRAJ9OSnvULf/ARrrVywoJDy0YS2AwQ98I37610=
 golang.org/x/sync v0.13.0/go.mod h1:1dzgHSNfp02xaA81J2MS99Qcpr2w7fw1gpm99rleRqA=
-golang.org/x/sys v0.0.0-20180830151530-49385e6e1522/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
 golang.org/x/sys v0.0.0-20180905080454-ebe1bf3edb33/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
 golang.org/x/sys v0.0.0-20180909124046-d0be0721c37e/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
 golang.org/x/sys v0.0.0-20190215142949-d0b11bdaac8a/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
@@ -396,10 +353,6 @@ golang.org/x/text v0.24.0/go.mod h1:L8rBsPeo2pSS+xqN0d5u2ikmjtmoJbDBT1b7nHvFCdU=
 golang.org/x/time v0.5.0 h1:o7cqy6amK/52YcAKIPlM3a+Fpj35zvRj2TP+e1xFSfk=
 golang.org/x/time v0.5.0/go.mod h1:3BpzKBy/shNhVucY/MWOyx10tF3SFh9QdLuxbVysPQM=
 golang.org/x/tools v0.0.0-20180917221912-90fa682c2a6e/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
-golang.org/x/tools v0.0.0-20190114222345-bf090417da8b/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
-golang.org/x/tools v0.0.0-20190226205152-f727befe758c/go.mod h1:9Yl7xja0Znq3iFh3HoIrodX9oNMXvdceNzlUR8zjMvY=
-golang.org/x/tools v0.0.0-20190311212946-11955173bddd/go.mod h1:LCzVGOaR6xXOjkQ3onu1FJEFr0SW1gC7cKk1uF8kGRs=
-golang.org/x/tools v0.0.0-20190524140312-2c0ae7006135/go.mod h1:RgjU9mgBXZiqYHBnxXauZ1Gv1EHHAz9KjViQ78xBX0Q=
 golang.org/x/tools v0.0.0-20191216052735-49a3e744a425/go.mod h1:TB2adYChydJhpapKDTa4BR/hXlZSLoq2Wpct/0txZ28=
 golang.org/x/tools v0.32.0 h1:Q7N1vhpkQv7ybVzLFtTjvQya2ewbwNDZzUgfXGqtMWU=
 golang.org/x/tools v0.32.0/go.mod h1:ZxrU41P/wAbZD8EDa6dDCa6XfpkhJ7HFMjHJXfBDu8s=
@@ -407,34 +360,16 @@ golang.org/x/xerrors v0.0.0-20191011141410-1b5146add898/go.mod h1:I/5z698sn9Ka8T
 golang.org/x/xerrors v0.0.0-20191204190536-9bdfabe68543/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
 golang.zx2c4.com/wintun v0.0.0-20230126152724-0fa3db229ce2 h1:B82qJJgjvYKsXS9jeunTOisW56dUokqW/FOteYJJ/yg=
 golang.zx2c4.com/wintun v0.0.0-20230126152724-0fa3db229ce2/go.mod h1:deeaetjYA+DHMHg+sMSMI58GrEteJUUzzw7en6TJQcI=
-google.golang.org/appengine v1.1.0/go.mod h1:EbEs0AVv82hx2wNQdGPgUI5lhzA/G0D9YwlJXL52JkM=
-google.golang.org/appengine v1.4.0/go.mod h1:xpcJRLb0r/rnEns0DIKYYv+WjYCduHsrkT7/EB5XEv4=
-google.golang.org/genproto v0.0.0-20180817151627-c66870c02cf8/go.mod h1:JiN7NxoALGmiZfu7CAH4rXhgtRTLTxftemlI0sWmxmc=
-google.golang.org/genproto v0.0.0-20190819201941-24fa4b261c55/go.mod h1:DMBHOl98Agz4BDEuKkezgsaosCRResVns1a3J2ZsMNc=
-google.golang.org/genproto v0.0.0-20200526211855-cb27e3aa2013/go.mod h1:NbSheEEYHJ7i3ixzK3sjbqSGDJWnxyFXZblF3eUsNvo=
 google.golang.org/genproto/googleapis/rpc v0.0.0-20240116215550-a9fa1716bcac h1:nUQEQmH/csSvFECKYRv6HWEyypysidKl2I6Qpsglq/0=
 google.golang.org/genproto/googleapis/rpc v0.0.0-20240116215550-a9fa1716bcac/go.mod h1:daQN87bsDqDoe316QbbvX60nMoJQa4r6Ds0ZuoAe5yA=
-google.golang.org/grpc v1.19.0/go.mod h1:mqu4LbDTu4XGKhr4mRzUsmM4RtVoemTSY81AxZiDr8c=
-google.golang.org/grpc v1.23.0/go.mod h1:Y5yQAOtifL1yxbo5wqy6BxZv8vAUGQwXBOALyacEbxg=
-google.golang.org/grpc v1.27.0/go.mod h1:qbnxyOmOxrQa7FizSgH+ReBfzJrCY1pSN7KXBS8abTk=
 google.golang.org/grpc v1.57.0 h1:kfzNeI/klCGD2YPMUlaGNT3pxvYfga7smW3Vth8Zsiw=
 google.golang.org/grpc v1.57.0/go.mod h1:Sd+9RMTACXwmub0zcNY2c4arhtrbBYD1AUHI/dt16Mo=
-google.golang.org/protobuf v0.0.0-20200109180630-ec00e32a8dfd/go.mod h1:DFci5gLYBciE7Vtevhsrf46CRTquxDuWsQurQQe4oz8=
-google.golang.org/protobuf v0.0.0-20200221191635-4d8936d0db64/go.mod h1:kwYJMbMJ01Woi6D6+Kah6886xMZcty6N08ah7+eCXa0=
-google.golang.org/protobuf v0.0.0-20200228230310-ab0ca4ff8a60/go.mod h1:cfTl7dwQJ+fmap5saPgwCLgHXTUD7jkjRqWcaiX5VyM=
-google.golang.org/protobuf v1.20.1-0.20200309200217-e05f789c0967/go.mod h1:A+miEFZTKqfCUM6K7xSMQL9OKL/b6hQv+e19PK+JZNE=
-google.golang.org/protobuf v1.21.0/go.mod h1:47Nbq4nVaFHyn7ilMalzfO3qCViNmqZ2kzikPIcrTAo=
-google.golang.org/protobuf v1.22.0/go.mod h1:EGpADcykh3NcUnDUJcl1+ZksZNG86OlYog2l/sGQquU=
-google.golang.org/protobuf v1.23.0/go.mod h1:EGpADcykh3NcUnDUJcl1+ZksZNG86OlYog2l/sGQquU=
-google.golang.org/protobuf v1.23.1-0.20200526195155-81db48ad09cc/go.mod h1:EGpADcykh3NcUnDUJcl1+ZksZNG86OlYog2l/sGQquU=
-google.golang.org/protobuf v1.25.0/go.mod h1:9JNX74DMeImyA3h4bdi1ymwjUzf21/xIlbajtzgsN7c=
 google.golang.org/protobuf v1.26.0-rc.1/go.mod h1:jlhhOSvTdKEhbULTjvd4ARK9grFBp09yW+WbY/TyQbw=
 google.golang.org/protobuf v1.26.0/go.mod h1:9q0QmTI4eRPtz6boOQmLYwt+qCgq0jsYwAQnmE0givc=
 google.golang.org/protobuf v1.34.1 h1:9ddQBjfCyZPOHPUiPxpYESBLc+T8P3E+Vo4IbKZgFWg=
 google.golang.org/protobuf v1.34.1/go.mod h1:c6P6GXX6sHbq/GpV6MGZEdwhWPcYBgnhAHhKbcUYpos=
 gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
 gopkg.in/check.v1 v1.0.0-20180628173108-788fd7840127/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
-gopkg.in/check.v1 v1.0.0-20190902080502-41f04d3bba15/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
 gopkg.in/check.v1 v1.0.0-20200227125254-8fa46927fb4f/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
 gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c h1:Hei/4ADfdWqJk1ZMxUNpqntNwaWcugrBjAiHlqqRiVk=
 gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c/go.mod h1:JHkPIbrfpd72SG/EVd6muEfDQjcINNoR0C8j2r3qZ4Q=
@@ -453,8 +388,6 @@ gopkg.in/yaml.v3 v3.0.1 h1:fxVm/GzAzEWqLHuvctI91KS9hhNmmWOoWu0XTYJS7CA=
 gopkg.in/yaml.v3 v3.0.1/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=
 gvisor.dev/gvisor v0.0.0-20240405191320-0878b34101b5 h1:DOUDfNS+CFMM46k18FRF5k/0yz5NhZYMiUQxf4xglIU=
 gvisor.dev/gvisor v0.0.0-20240405191320-0878b34101b5/go.mod h1:NQHVAzMwvZ+Qe3ElSiHmq9RUm1MdNHpUZ52fiEqvn+0=
-honnef.co/go/tools v0.0.0-20190102054323-c2f93a96b099/go.mod h1:rf3lG4BRIbNafJWhAfAdb/ePZxsR/4RtNHQocxwk9r4=
-honnef.co/go/tools v0.0.0-20190523083050-ea95bdfd59fc/go.mod h1:rf3lG4BRIbNafJWhAfAdb/ePZxsR/4RtNHQocxwk9r4=
 howett.net/plist v1.0.0 h1:7CrbWYbPPO/PyNy38b2EB/+gYbjCe2DXBxgtOOZbSQM=
 howett.net/plist v1.0.0/go.mod h1:lqaXoTrLY4hg8tnEzNru53gicrbv7rrk+2xJA/7hw9g=
 nullprogram.com/x/optparse v1.0.0/go.mod h1:KdyPE+Igbe0jQUrVfMqDMeJQIJZEuyV7pjYmp6pbG50=
--- a/internal/builtin/utils.go
+++ b/internal/builtin/utils.go
@@ -6,9 +6,13 @@ import (
 	"context"
 	"crypto/hmac"
 	"crypto/sha256"
+	"encoding/base64"
 	"encoding/csv"
 	builtinJSON "encoding/json"
 	"fmt"
+	"image"
+	"image/jpeg"
+	"image/png"
 	"math"
 	"math/rand"
 	"net"
@@ -26,6 +30,7 @@ import (

 	"github.com/httprunner/httprunner/v5/code"
 	"github.com/httprunner/httprunner/v5/internal/json"
+	"github.com/httprunner/httprunner/v5/uixt/types"
 )

 func Dump2JSON(data interface{}, path string) error {
@@ -484,3 +489,41 @@ func RunCommandWithCallback(cmdName string, args []string, callback LineCallback
 		return err
 	}
 }
+
+// LoadImage loads image file and returns base64 encoded string and image size
+func LoadImage(imagePath string) (base64Str string, size types.Size, err error) {
+	// Read the image file
+	imageFile, err := os.Open(imagePath)
+	if err != nil {
+		return "", types.Size{}, fmt.Errorf("failed to open image file: %w", err)
+	}
+	defer imageFile.Close()
+
+	// Decode the image to get its resolution
+	imageData, format, err := image.Decode(imageFile)
+	if err != nil {
+		return "", types.Size{}, fmt.Errorf("failed to decode image: %w", err)
+	}
+
+	// Get the resolution of the image
+	width := imageData.Bounds().Dx()
+	height := imageData.Bounds().Dy()
+	size = types.Size{Width: width, Height: height}
+
+	// Convert image to base64
+	buf := new(bytes.Buffer)
+	if format == "jpeg" || format == "jpg" {
+		if err := jpeg.Encode(buf, imageData, nil); err != nil {
+			return "", types.Size{}, fmt.Errorf("failed to encode image to buffer: %w", err)
+		}
+	} else {
+		// default use png
+		if err := png.Encode(buf, imageData); err != nil {
+			return "", types.Size{}, fmt.Errorf("failed to encode image to buffer: %w", err)
+		}
+	}
+	base64Str = fmt.Sprintf("data:image/%s;base64,%s", format,
+		base64.StdEncoding.EncodeToString(buf.Bytes()))
+
+	return base64Str, size, nil
+}
--- a/internal/version/VERSION
+++ b/internal/version/VERSION
@@ -1 +1 @@
-v5.0.0-beta-2504292008
+v5.0.0-beta-2505071715
--- a/pkg/gadb/device.go
+++ b/pkg/gadb/device.go
@@ -536,7 +536,7 @@ func (d *Device) List(remotePath string) (devFileInfos []DeviceFileInfo, err err
 }

 func (d *Device) PushFile(localPath, remotePath string, modification ...time.Time) (err error) {
-	localFile, err := os.OpenFile(localPath, os.O_RDONLY, 0o600)
+	localFile, err := os.Open(localPath)
 	if err != nil {
 		return err
 	}
@@ -645,7 +645,7 @@ func (d *Device) installViaABBExec(apk io.ReadSeeker, args ...string) (raw []byt
 }

 func (d *Device) InstallAPK(apkPath string, args ...string) (string, error) {
-	apkFile, err := os.OpenFile(apkPath, os.O_RDONLY, 0o600)
+	apkFile, err := os.Open(apkPath)
 	if err != nil {
 		return "", errors.Wrap(err, fmt.Sprintf("open apk file %s failed", apkPath))
 	}
--- a/runner.go
+++ b/runner.go
@@ -28,7 +28,7 @@ import (
 	"github.com/httprunner/httprunner/v5/internal/sdk"
 	"github.com/httprunner/httprunner/v5/internal/version"
 	"github.com/httprunner/httprunner/v5/uixt"
-	"github.com/httprunner/httprunner/v5/uixt/ai"
+	"github.com/httprunner/httprunner/v5/uixt/option"
 )

 // Run starts to run testcase with default configs.
@@ -419,15 +419,15 @@ func (r *CaseRunner) parseConfig() (parsedConfig *TConfig, err error) {
 	r.parametersIterator = parametersIterator

 	// ai options
-	aiOpts := []ai.AIServiceOption{}
+	aiOpts := []option.AIServiceOption{}
 	if parsedConfig.LLMService != "" {
-		aiOpts = append(aiOpts, ai.WithLLMService(parsedConfig.LLMService))
+		aiOpts = append(aiOpts, option.WithLLMService(option.LLMServiceType(parsedConfig.LLMService)))
 	}
 	if parsedConfig.CVService == "" {
 		// default to vedem
-		parsedConfig.CVService = ai.CVServiceTypeVEDEM
+		parsedConfig.CVService = option.CVServiceTypeVEDEM
 	}
-	aiOpts = append(aiOpts, ai.WithCVService(parsedConfig.CVService))
+	aiOpts = append(aiOpts, option.WithCVService(parsedConfig.CVService))

 	// parse android devices config
 	for _, androidDeviceOptions := range parsedConfig.Android {
@@ -446,7 +446,10 @@ func (r *CaseRunner) parseConfig() (parsedConfig *TConfig, err error) {
 			return nil, errors.Wrap(err, "init android driver failed")
 		}

-		driverExt := uixt.NewXTDriver(driver, aiOpts...)
+		driverExt, err := uixt.NewXTDriver(driver, aiOpts...)
+		if err != nil {
+			return nil, errors.Wrap(err, "init android XTDriver failed")
+		}
 		r.uixtDrivers[androidDeviceOptions.SerialNumber] = driverExt
 	}
 	// parse iOS devices config
@@ -466,7 +469,10 @@ func (r *CaseRunner) parseConfig() (parsedConfig *TConfig, err error) {
 			return nil, errors.Wrap(err, "init ios driver failed")
 		}

-		driverExt := uixt.NewXTDriver(driver, aiOpts...)
+		driverExt, err := uixt.NewXTDriver(driver, aiOpts...)
+		if err != nil {
+			return nil, errors.Wrap(err, "init ios XTDriver failed")
+		}
 		r.uixtDrivers[iosDeviceOptions.UDID] = driverExt
 	}
 	// parse harmony devices config
@@ -486,7 +492,10 @@ func (r *CaseRunner) parseConfig() (parsedConfig *TConfig, err error) {
 			return nil, errors.Wrap(err, "init harmony driver failed")
 		}

-		driverExt := uixt.NewXTDriver(driver, aiOpts...)
+		driverExt, err := uixt.NewXTDriver(driver, aiOpts...)
+		if err != nil {
+			return nil, errors.Wrap(err, "init harmony XTDriver failed")
+		}
 		r.uixtDrivers[harmonyDeviceOptions.ConnectKey] = driverExt
 	}

--- a/server/context.go
+++ b/server/context.go
@@ -10,7 +10,6 @@ import (

 	"github.com/httprunner/httprunner/v5/code"
 	"github.com/httprunner/httprunner/v5/uixt"
-	"github.com/httprunner/httprunner/v5/uixt/ai"
 	"github.com/httprunner/httprunner/v5/uixt/option"
 )

@@ -33,8 +32,12 @@ func (r *Router) GetDriver(c *gin.Context) (driverExt *uixt.XTDriver, err error)
 		return
 	}

-	driverExt = uixt.NewXTDriver(driver,
-		ai.WithCVService(ai.CVServiceTypeVEDEM))
+	driverExt, err = uixt.NewXTDriver(driver,
+		option.WithCVService(option.CVServiceTypeVEDEM))
+	if err != nil {
+		RenderErrorInitDriver(c, err)
+		return
+	}
 	c.Set("driver", driverExt)
 	return driverExt, nil
 }
--- a/summary.go
+++ b/summary.go
@@ -95,7 +95,7 @@ func (s *Summary) GenHTMLReport() error {
 	}

 	reportPath := filepath.Join(reportsDir, "report.html")
-	file, err := os.OpenFile(reportPath, os.O_WRONLY|os.O_CREATE, 0o600)
+	file, err := os.Open(reportPath)
 	if err != nil {
 		log.Error().Err(err).Msg("open file failed")
 		return err
--- a/uixt/ai/ai.go
+++ b/uixt/ai/ai.go
@@ -3,55 +3,14 @@ package ai
 import (
 	"context"
 	"os"
+	"time"

-	"github.com/rs/zerolog/log"
-
+	"github.com/cloudwego/eino-ext/components/model/openai"
 	"github.com/httprunner/httprunner/v5/code"
-)
-
-func NewAIService(opts ...AIServiceOption) *AIServices {
-	services := &AIServices{}
-	for _, option := range opts {
-		option(services)
-	}
-	return services
-}
-
-type AIServices struct {
-	ICVService
-	ILLMService
-}
-
-type AIServiceOption func(*AIServices)
-
-type CVServiceType string
-
-const (
-	CVServiceTypeVEDEM  CVServiceType = "vedem"
-	CVServiceTypeOpenCV CVServiceType = "opencv"
-)
-
-func WithCVService(service CVServiceType) AIServiceOption {
-	return func(opts *AIServices) {
-		if service == CVServiceTypeVEDEM {
-			var err error
-			opts.ICVService, err = NewVEDEMImageService()
-			if err != nil {
-				log.Error().Err(err).Msg("init vedem image service failed")
-				os.Exit(code.GetErrorCode(err))
-			}
-		}
-	}
-}
-
-type LLMServiceType string
-
-const (
-	LLMServiceTypeUITARS     LLMServiceType = "ui-tars"
-	LLMServiceTypeGPT4o      LLMServiceType = "gpt-4o"
-	LLMServiceTypeGPT4Vision LLMServiceType = "gpt-4-vision"
-	LLMServiceTypeQwenVL     LLMServiceType = "qwen-vl"
-	LLMServiceTypeDeepSeekV3 LLMServiceType = "deepseek-v3"
+	"github.com/httprunner/httprunner/v5/internal/config"
+	"github.com/httprunner/httprunner/v5/uixt/option"
+	"github.com/pkg/errors"
+	"github.com/rs/zerolog/log"
 )

 // ILLMService 定义了 LLM 服务接口，包括规划和断言功能
@@ -60,38 +19,29 @@ type ILLMService interface {
 	Assert(opts *AssertOptions) (*AssertionResponse, error)
 }

-func WithLLMService(modelType LLMServiceType) AIServiceOption {
-	return func(opts *AIServices) {
-		// init planner
-		var planner IPlanner
-		var err error
-		switch modelType {
-		case LLMServiceTypeGPT4o:
-			// TODO: implement gpt-4o planner and asserter
-			planner, err = NewPlanner(context.Background())
-		case LLMServiceTypeUITARS:
-			planner, err = NewUITarsPlanner(context.Background())
-		}
-		if err != nil {
-			log.Error().Err(err).Msgf("init %s planner failed", modelType)
-			os.Exit(code.GetErrorCode(err))
-		}
-
-		// init asserter
-		asserter, err := NewAsserter(context.Background(), modelType)
-		if err != nil {
-			log.Error().Err(err).Msgf("init %s asserter failed", modelType)
-			os.Exit(code.GetErrorCode(err))
-		}
-
-		opts.ILLMService = &combinedLLMService{
-			planner:  planner,
-			asserter: asserter,
-		}
+func NewLLMService(modelType option.LLMServiceType) (ILLMService, error) {
+	modelConfig, err := GetModelConfig(modelType)
+	if err != nil {
+		return nil, err
 	}
+
+	planner, err := NewPlanner(context.Background(), modelConfig)
+	if err != nil {
+		return nil, err
+	}
+	asserter, err := NewAsserter(context.Background(), modelConfig)
+	if err != nil {
+		return nil, err
+	}
+
+	return &combinedLLMService{
+		planner:  planner,
+		asserter: asserter,
+	}, nil
 }

 // combinedLLMService 实现了 ILLMService 接口，组合了规划和断言功能
+// ⭐️支持采用不同的模型服务进行规划和断言
 type combinedLLMService struct {
 	planner  IPlanner  // 提供规划功能
 	asserter IAsserter // 提供断言功能
@@ -106,3 +56,72 @@ func (c *combinedLLMService) Call(opts *PlanningOptions) (*PlanningResult, error
 func (c *combinedLLMService) Assert(opts *AssertOptions) (*AssertionResponse, error) {
 	return c.asserter.Assert(opts)
 }
+
+// LLM model config env variables
+const (
+	EnvOpenAIBaseURL = "OPENAI_BASE_URL"
+	EnvOpenAIAPIKey  = "OPENAI_API_KEY"
+	EnvModelName     = "LLM_MODEL_NAME"
+)
+
+const (
+	defaultTimeout = 30 * time.Second
+)
+
+type ModelConfig struct {
+	*openai.ChatModelConfig
+	ModelType option.LLMServiceType
+}
+
+// GetModelConfig get OpenAI config
+func GetModelConfig(modelType option.LLMServiceType) (*ModelConfig, error) {
+	if err := config.LoadEnv(); err != nil {
+		return nil, errors.Wrap(code.LoadEnvError, err.Error())
+	}
+
+	openaiBaseURL := os.Getenv(EnvOpenAIBaseURL)
+	if openaiBaseURL == "" {
+		return nil, errors.Wrapf(code.LLMEnvMissedError,
+			"env %s missed", EnvOpenAIBaseURL)
+	}
+	openaiAPIKey := os.Getenv(EnvOpenAIAPIKey)
+	if openaiAPIKey == "" {
+		return nil, errors.Wrapf(code.LLMEnvMissedError,
+			"env %s missed", EnvOpenAIAPIKey)
+	}
+	modelName := os.Getenv(EnvModelName)
+	if modelName == "" {
+		return nil, errors.Wrapf(code.LLMEnvMissedError,
+			"env %s missed", EnvModelName)
+	}
+
+	temperature := float32(0.01)
+	modelConfig := &openai.ChatModelConfig{
+		BaseURL:     openaiBaseURL,
+		APIKey:      openaiAPIKey,
+		Model:       modelName,
+		Timeout:     defaultTimeout,
+		Temperature: &temperature,
+	}
+
+	// log config info
+	log.Info().Str("model", modelConfig.Model).
+		Str("baseURL", modelConfig.BaseURL).
+		Str("apiKey", maskAPIKey(modelConfig.APIKey)).
+		Str("timeout", defaultTimeout.String()).
+		Msg("get model config")
+
+	return &ModelConfig{
+		ChatModelConfig: modelConfig,
+		ModelType:       modelType,
+	}, nil
+}
+
+// maskAPIKey masks the API key
+func maskAPIKey(key string) string {
+	if len(key) <= 8 {
+		return "******"
+	}
+
+	return key[:4] + "******" + key[len(key)-4:]
+}
--- a/uixt/ai/ai_ark.go
+++ b/uixt/ai/ai_ark.go
@@ -1,62 +0,0 @@
-package ai
-
-import (
-	"os"
-
-	"github.com/cloudwego/eino-ext/components/model/ark"
-	"github.com/httprunner/httprunner/v5/code"
-	"github.com/httprunner/httprunner/v5/internal/config"
-	"github.com/pkg/errors"
-	"github.com/rs/zerolog/log"
-)
-
-const (
-	EnvArkBaseURL = "ARK_BASE_URL"
-	EnvArkAPIKey  = "ARK_API_KEY"
-	EnvArkModelID = "ARK_MODEL_ID"
-)
-
-func GetArkModelConfig() (*ark.ChatModelConfig, error) {
-	if err := config.LoadEnv(); err != nil {
-		return nil, errors.Wrap(code.LoadEnvError, err.Error())
-	}
-
-	arkBaseURL := os.Getenv(EnvArkBaseURL)
-	arkAPIKey := os.Getenv(EnvArkAPIKey)
-	if arkAPIKey == "" {
-		return nil, errors.Wrapf(code.LLMEnvMissedError,
-			"env %s missed", EnvArkAPIKey)
-	}
-	modelName := os.Getenv(EnvArkModelID)
-	if modelName == "" {
-		return nil, errors.Wrapf(code.LLMEnvMissedError,
-			"env %s missed", EnvArkModelID)
-	}
-	timeout := defaultTimeout
-
-	// https://www.volcengine.com/docs/82379/1494384?redirect=1
-	temperature := float32(0.01) // [0, 2] 采样温度。控制了生成文本时对每个候选词的概率分布进行平滑的程度。
-	// topP := float32(0.7)           // [0, 1] 核采样概率阈值。模型会考虑概率质量在 top_p 内的 token 结果。
-	// maxTokens := int(4096)         // 模型可以生成的最大 token 数量。输入 token 和输出 token 的总长度还受模型的上下文长度限制。
-	// frequencyPenalty := float32(0) // [-2, 2] 频率惩罚系数。如果值为正，会根据新 token 在文本中的出现频率对其进行惩罚，从而降低模型逐字重复的可能性。
-
-	modelConfig := &ark.ChatModelConfig{
-		BaseURL:     arkBaseURL,
-		APIKey:      arkAPIKey,
-		Model:       modelName,
-		Timeout:     &timeout,
-		Temperature: &temperature,
-		// TopP:             &topP,
-		// MaxTokens:        &maxTokens,
-		// FrequencyPenalty: &frequencyPenalty,
-	}
-
-	// log config info
-	log.Info().Str("model", modelConfig.Model).
-		Str("baseURL", modelConfig.BaseURL).
-		Str("apiKey", maskAPIKey(modelConfig.APIKey)).
-		Str("timeout", defaultTimeout.String()).
-		Msg("get model config")
-
-	return modelConfig, nil
-}
--- a/uixt/ai/ai_openai.go
+++ b/uixt/ai/ai_openai.go
@@ -1,79 +0,0 @@
-package ai
-
-import (
-	"os"
-
-	"github.com/cloudwego/eino-ext/components/model/openai"
-	openai2 "github.com/cloudwego/eino-ext/libs/acl/openai"
-	"github.com/getkin/kin-openapi/openapi3gen"
-	"github.com/httprunner/httprunner/v5/code"
-	"github.com/httprunner/httprunner/v5/internal/config"
-	"github.com/pkg/errors"
-	"github.com/rs/zerolog/log"
-)
-
-const (
-	EnvOpenAIBaseURL = "OPENAI_BASE_URL"
-	EnvOpenAIAPIKey  = "OPENAI_API_KEY"
-	EnvModelName     = "LLM_MODEL_NAME"
-)
-
-// GetOpenAIModelConfig get OpenAI config
-func GetOpenAIModelConfig() (*openai.ChatModelConfig, error) {
-	if err := config.LoadEnv(); err != nil {
-		return nil, errors.Wrap(code.LoadEnvError, err.Error())
-	}
-
-	openaiBaseURL := os.Getenv(EnvOpenAIBaseURL)
-	if openaiBaseURL == "" {
-		return nil, errors.Wrapf(code.LLMEnvMissedError,
-			"env %s missed", EnvOpenAIBaseURL)
-	}
-	openaiAPIKey := os.Getenv(EnvOpenAIAPIKey)
-	if openaiAPIKey == "" {
-		return nil, errors.Wrapf(code.LLMEnvMissedError,
-			"env %s missed", EnvOpenAIAPIKey)
-	}
-	modelName := os.Getenv(EnvModelName)
-	if modelName == "" {
-		return nil, errors.Wrapf(code.LLMEnvMissedError,
-			"env %s missed", EnvModelName)
-	}
-
-	type OutputFormat struct {
-		Thought string `json:"thought"`
-		Action  string `json:"action"`
-		Error   string `json:"error,omitempty"`
-	}
-	outputFormatSchema, err := openapi3gen.NewSchemaRefForValue(&OutputFormat{}, nil)
-	if err != nil {
-		return nil, err
-	}
-
-	modelConfig := &openai.ChatModelConfig{
-		BaseURL: openaiBaseURL,
-		APIKey:  openaiAPIKey,
-		Model:   modelName,
-		Timeout: defaultTimeout,
-		// set structured response format
-		// https://github.com/cloudwego/eino-ext/blob/main/components/model/openai/examples/structured/structured.go
-		ResponseFormat: &openai2.ChatCompletionResponseFormat{
-			Type: openai2.ChatCompletionResponseFormatTypeJSONSchema,
-			JSONSchema: &openai2.ChatCompletionResponseFormatJSONSchema{
-				Name:        "thought_and_action",
-				Description: "data that describes planning thought and action",
-				Schema:      outputFormatSchema.Value,
-				Strict:      false,
-			},
-		},
-	}
-
-	// log config info
-	log.Info().Str("model", modelConfig.Model).
-		Str("baseURL", modelConfig.BaseURL).
-		Str("apiKey", maskAPIKey(modelConfig.APIKey)).
-		Str("timeout", defaultTimeout.String()).
-		Msg("get model config")
-
-	return modelConfig, nil
-}
--- a/uixt/ai/ai_test.go
+++ b/uixt/ai/ai_test.go
@@ -1,13 +0,0 @@
-package ai
-
-import (
-	"testing"
-)
-
-func TestOption(t *testing.T) {
-	options := NewAIService(
-		WithCVService(CVServiceTypeOpenCV),
-		WithLLMService(LLMServiceTypeUITARS),
-	)
-	t.Log(options)
-}
--- a/uixt/ai/asserter.go
+++ b/uixt/ai/asserter.go
@@ -7,12 +7,14 @@ import (
 	"strings"
 	"time"

-	"github.com/cloudwego/eino-ext/components/model/ark"
 	"github.com/cloudwego/eino-ext/components/model/openai"
+	openai2 "github.com/cloudwego/eino-ext/libs/acl/openai"
 	"github.com/cloudwego/eino/components/model"
 	"github.com/cloudwego/eino/schema"
+	"github.com/getkin/kin-openapi/openapi3gen"
 	"github.com/httprunner/httprunner/v5/code"
 	"github.com/httprunner/httprunner/v5/internal/json"
+	"github.com/httprunner/httprunner/v5/uixt/option"
 	"github.com/httprunner/httprunner/v5/uixt/types"
 	"github.com/pkg/errors"
 	"github.com/rs/zerolog/log"
@@ -39,54 +41,57 @@ type AssertionResponse struct {
 // Asserter handles assertion using different AI models
 type Asserter struct {
 	ctx          context.Context
+	modelConfig  *ModelConfig
 	model        model.ToolCallingChatModel
 	systemPrompt string
 	history      ConversationHistory
-	modelType    LLMServiceType
 }

 // NewAsserter creates a new Asserter instance
-func NewAsserter(ctx context.Context, modelType LLMServiceType) (*Asserter, error) {
+func NewAsserter(ctx context.Context, modelConfig *ModelConfig) (*Asserter, error) {
 	asserter := &Asserter{
 		ctx:          ctx,
-		modelType:    modelType,
-		systemPrompt: getAssertionSystemPrompt(modelType),
+		modelConfig:  modelConfig,
+		systemPrompt: defaultAssertionPrompt,
 	}

-	switch modelType {
-	case LLMServiceTypeUITARS:
-		config, err := GetArkModelConfig()
-		if err != nil {
-			return nil, err
+	if modelConfig.ModelType == option.LLMServiceTypeUITARS {
+		asserter.systemPrompt += "\n\n" + uiTarsAssertionResponseFormat
+	} else if modelConfig.ModelType == option.LLMServiceTypeGPT {
+		// define output format
+		type OutputFormat struct {
+			Thought string `json:"thought"`
+			Pass    bool   `json:"pass"`
+			Error   string `json:"error,omitempty"`
 		}
-		asserter.model, err = ark.NewChatModel(ctx, config)
+		outputFormatSchema, err := openapi3gen.NewSchemaRefForValue(&OutputFormat{}, nil)
 		if err != nil {
-			return nil, err
+			return nil, errors.Wrap(code.LLMPrepareRequestError, err.Error())
 		}
-	case LLMServiceTypeGPT4Vision, LLMServiceTypeGPT4o:
-		config, err := GetOpenAIModelConfig()
-		if err != nil {
-			return nil, err
+		// set structured response format
+		// https://github.com/cloudwego/eino-ext/blob/main/components/model/openai/examples/structured/structured.go
+		modelConfig.ChatModelConfig.ResponseFormat = &openai2.ChatCompletionResponseFormat{
+			Type: openai2.ChatCompletionResponseFormatTypeJSONSchema,
+			JSONSchema: &openai2.ChatCompletionResponseFormatJSONSchema{
+				Name:        "assertion_result",
+				Description: "data that describes assertion result",
+				Schema:      outputFormatSchema.Value,
+				Strict:      false,
+			},
 		}
-		asserter.model, err = openai.NewChatModel(ctx, config)
-		if err != nil {
-			return nil, err
-		}
-	default:
-		return nil, errors.New("not supported model type for asserter")
+	} else {
+		asserter.systemPrompt += "\n\n" + defaultAssertionResponseJsonFormat
+	}
+
+	var err error
+	asserter.model, err = openai.NewChatModel(ctx, modelConfig.ChatModelConfig)
+	if err != nil {
+		return nil, errors.Wrap(code.LLMPrepareRequestError, err.Error())
 	}

 	return asserter, nil
 }

-// getAssertionSystemPrompt returns the appropriate system prompt for the given model type
-func getAssertionSystemPrompt(modelType LLMServiceType) string {
-	if modelType == LLMServiceTypeUITARS {
-		return defaultAssertionPrompt + "\n\n" + uiTarsAssertionResponseFormat
-	}
-	return defaultAssertionPrompt + "\n\n" + defaultAssertionResponseJsonFormat
-}
-
 // Assert performs the assertion check on the screenshot
 func (a *Asserter) Assert(opts *AssertOptions) (*AssertionResponse, error) {
 	// Validate input parameters
@@ -133,7 +138,7 @@ Here is the assertion. Please tell whether it is truthy according to the screens
 	startTime := time.Now()
 	resp, err := a.model.Generate(a.ctx, a.history)
 	log.Info().Float64("elapsed(s)", time.Since(startTime).Seconds()).
-		Str("model", string(a.modelType)).Msg("call model service for assertion")
+		Str("model", string(a.modelConfig.ModelType)).Msg("call model service for assertion")
 	if err != nil {
 		return nil, errors.Wrap(code.LLMRequestServiceError, err.Error())
 	}
--- a/uixt/ai/asserter_test.go
+++ b/uixt/ai/asserter_test.go
@@ -1,23 +1,27 @@
 package ai

 import (
+	"context"
 	"testing"

+	"github.com/httprunner/httprunner/v5/internal/builtin"
+	"github.com/httprunner/httprunner/v5/uixt/option"
 	"github.com/httprunner/httprunner/v5/uixt/types"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 )

-func createAIService(t *testing.T) *AIServices {
-	aiService := NewAIService(WithLLMService(LLMServiceTypeUITARS))
-	require.NotNil(t, aiService)
-	require.NotNil(t, aiService.ILLMService)
-	return aiService
+func createAsserter(t *testing.T) *Asserter {
+	modelConfig, err := GetModelConfig(option.LLMServiceTypeUITARS)
+	require.NoError(t, err)
+	asserter, err := NewAsserter(context.Background(), modelConfig)
+	require.NoError(t, err)
+	return asserter
 }

 // 测试有效断言
 func TestValidAssertions(t *testing.T) {
-	aiService := createAIService(t)
+	asserter := createAsserter(t)

 	testCases := []struct {
 		name       string
@@ -33,7 +37,7 @@ func TestValidAssertions(t *testing.T) {
 		},
 		{
 			name:       "深度思考功能未开启",
-			assertion:  "输入框下方的「深度思考」文字是灰色的",
+			assertion:  "输入框下方的「深度思考」文字不是蓝色的",
 			imagePath:  "testdata/deepseek_think_off.png",
 			expectPass: true,
 		},
@@ -47,10 +51,10 @@ func TestValidAssertions(t *testing.T) {

 	for _, tc := range testCases {
 		t.Run(tc.name, func(t *testing.T) {
-			imageBase64, size, err := loadImage(tc.imagePath)
+			imageBase64, size, err := builtin.LoadImage(tc.imagePath)
 			require.NoError(t, err)

-			result, err := aiService.ILLMService.Assert(&AssertOptions{
+			result, err := asserter.Assert(&AssertOptions{
 				Assertion:  tc.assertion,
 				Screenshot: imageBase64,
 				Size:       size,
@@ -58,14 +62,13 @@ func TestValidAssertions(t *testing.T) {
 			require.NoError(t, err)
 			assert.NotNil(t, result)
 			assert.Equal(t, tc.expectPass, result.Pass)
-			assert.NotEmpty(t, result.Thought)
 		})
 	}
 }

 // 测试无效参数
 func TestInvalidParameters(t *testing.T) {
-	aiService := createAIService(t)
+	asserter := createAsserter(t)
 	testCases := []struct {
 		name          string
 		assertion     string
@@ -91,7 +94,7 @@ func TestInvalidParameters(t *testing.T) {

 	for _, tc := range testCases {
 		t.Run(tc.name, func(t *testing.T) {
-			_, err := aiService.ILLMService.Assert(&AssertOptions{
+			_, err := asserter.Assert(&AssertOptions{
 				Assertion:  tc.assertion,
 				Screenshot: tc.screenshot,
 				Size:       tc.size,
--- a/uixt/ai/cv.go
+++ b/uixt/ai/cv.go
@@ -22,6 +22,13 @@ type ICVService interface {
 	ReadFromPath(imagePath string, opts ...option.ActionOption) (*CVResult, error)
 }

+func NewCVService(modelType option.CVServiceType) (ICVService, error) {
+	if modelType == option.CVServiceTypeVEDEM {
+		return NewVEDEMImageService()
+	}
+	return nil, errors.New("invalid cv service type")
+}
+
 type CVResult struct {
 	URL       string     `json:"url,omitempty"`       // image uploaded url
 	OCRResult OCRResults `json:"ocrResult,omitempty"` // OCR texts
--- a/uixt/ai/cv_vedem_test.go
+++ b/uixt/ai/cv_vedem_test.go
@@ -19,9 +19,8 @@ func TestGetImageFromBuffer(t *testing.T) {
 	buf := new(bytes.Buffer)
 	buf.Read(file)

-	service := NewAIService(
-		WithCVService(CVServiceTypeVEDEM),
-	)
+	service, err := NewVEDEMImageService()
+	require.Nil(t, err)
 	cvResult, err := service.ReadFromBuffer(buf)
 	assert.Nil(t, err)
 	fmt.Println(fmt.Sprintf("cvResult: %v", cvResult))
@@ -29,9 +28,8 @@ func TestGetImageFromBuffer(t *testing.T) {

 func TestGetImageFromPath(t *testing.T) {
 	imagePath := "/Users/debugtalk/Downloads/s1.png"
-	service := NewAIService(
-		WithCVService(CVServiceTypeVEDEM),
-	)
+	service, err := NewVEDEMImageService()
+	require.Nil(t, err)
 	cvResult, err := service.ReadFromPath(imagePath)
 	assert.Nil(t, err)
 	fmt.Println(fmt.Sprintf("cvResult: %v", cvResult))
--- a/uixt/ai/planner.go
+++ b/uixt/ai/planner.go
@@ -1,21 +1,17 @@
 package ai

 import (
-	"bytes"
-	"encoding/base64"
-	"fmt"
-	"image"
-	"image/color"
-	"image/draw"
-	"image/png"
-	"os"
-	"strings"
+	"context"
 	"time"

+	"github.com/cloudwego/eino-ext/components/model/openai"
+	"github.com/cloudwego/eino/components/model"
 	"github.com/cloudwego/eino/schema"
 	"github.com/httprunner/httprunner/v5/code"
+	"github.com/httprunner/httprunner/v5/uixt/option"
 	"github.com/httprunner/httprunner/v5/uixt/types"
 	"github.com/pkg/errors"
+	"github.com/rs/zerolog/log"
 )

 type IPlanner interface {
@@ -36,30 +32,110 @@ type PlanningResult struct {
 	Error         string         `json:"error,omitempty"`
 }

-// ParsedAction represents a parsed action from the VLM response
-type ParsedAction struct {
-	ActionType   ActionType             `json:"actionType"`
-	ActionInputs map[string]interface{} `json:"actionInputs"`
-	Thought      string                 `json:"thought"`
+func NewPlanner(ctx context.Context, modelConfig *ModelConfig) (*Planner, error) {
+	planner := &Planner{
+		ctx:         ctx,
+		modelConfig: modelConfig,
+	}
+
+	if modelConfig.ModelType == option.LLMServiceTypeUITARS {
+		planner.systemPrompt = uiTarsPlanningPrompt
+	} else {
+		planner.systemPrompt = defaultPlanningResponseJsonFormat
+	}
+
+	var err error
+	planner.model, err = openai.NewChatModel(ctx, modelConfig.ChatModelConfig)
+	if err != nil {
+		return nil, errors.Wrap(code.LLMPrepareRequestError, err.Error())
+	}
+
+	return planner, nil
 }

-type ActionType string
+type Planner struct {
+	ctx          context.Context
+	modelConfig  *ModelConfig
+	model        model.ToolCallingChatModel
+	systemPrompt string
+	history      ConversationHistory
+}

-const (
-	ActionTypeClick    ActionType = "click"
-	ActionTypeTap      ActionType = "tap"
-	ActionTypeDrag     ActionType = "drag"
-	ActionTypeSwipe    ActionType = "swipe"
-	ActionTypeWait     ActionType = "wait"
-	ActionTypeFinished ActionType = "finished"
-	ActionTypeCallUser ActionType = "call_user"
-	ActionTypeType     ActionType = "type"
-	ActionTypeScroll   ActionType = "scroll"
-)
+// Call performs UI planning using Vision Language Model
+func (p *Planner) Call(opts *PlanningOptions) (*PlanningResult, error) {
+	// validate input parameters
+	if err := validatePlanningInput(opts); err != nil {
+		return nil, errors.Wrap(err, "validate planning parameters failed")
+	}

-const (
-	defaultTimeout = 30 * time.Second
-)
+	// prepare prompt
+	if len(p.history) == 0 {
+		// add system message
+		p.history = ConversationHistory{
+			{
+				Role:    schema.System,
+				Content: p.systemPrompt + opts.UserInstruction,
+			},
+		}
+	}
+	// append user image message
+	p.history.Append(opts.Message)
+
+	// call model service, generate response
+	logRequest(p.history)
+	startTime := time.Now()
+	resp, err := p.model.Generate(p.ctx, p.history)
+	log.Info().Float64("elapsed(s)", time.Since(startTime).Seconds()).
+		Str("model", string(p.modelConfig.ModelType)).Msg("call model service")
+	if err != nil {
+		return nil, errors.Wrap(code.LLMRequestServiceError, err.Error())
+	}
+	logResponse(resp)
+
+	// parse result
+	result, err := p.parseResult(resp, opts.Size)
+	if err != nil {
+		return nil, errors.Wrap(code.LLMParsePlanningResponseError, err.Error())
+	}
+
+	// append assistant message
+	p.history.Append(&schema.Message{
+		Role:    schema.Assistant,
+		Content: result.ActionSummary,
+	})
+
+	return result, nil
+}
+
+func (p *Planner) parseResult(msg *schema.Message, size types.Size) (*PlanningResult, error) {
+	var parseActions []ParsedAction
+	var err error
+	if p.modelConfig.ModelType == option.LLMServiceTypeUITARS {
+		// parse Thought/Action format from UI-TARS
+		parseActions, err = parseThoughtAction(msg.Content)
+		if err != nil {
+			return nil, err
+		}
+	} else {
+		// parse JSON format, from VLM like openai/gpt-4o
+		parseActions, err = parseJSON(msg.Content)
+		if err != nil {
+			return nil, err
+		}
+	}
+
+	// process response
+	result, err := processVLMResponse(parseActions, size)
+	if err != nil {
+		return nil, errors.Wrap(err, "process VLM response failed")
+	}
+
+	log.Info().
+		Interface("summary", result.ActionSummary).
+		Interface("actions", result.NextActions).
+		Msg("get VLM planning result")
+	return result, nil
+}

 func validatePlanningInput(opts *PlanningOptions) error {
 	if opts.UserInstruction == "" {
@@ -83,79 +159,3 @@ func validatePlanningInput(opts *PlanningOptions) error {

 	return nil
 }
-
-// SavePositionImg saves an image with position markers
-func SavePositionImg(params struct {
-	InputImgBase64 string
-	Rect           struct {
-		X float64
-		Y float64
-	}
-	OutputPath string
-}) error {
-	// 解码Base64图像
-	imgData := params.InputImgBase64
-	// 如果包含了数据URL前缀，去掉它
-	if strings.HasPrefix(imgData, "data:image/") {
-		parts := strings.Split(imgData, ",")
-		if len(parts) > 1 {
-			imgData = parts[1]
-		}
-	}
-
-	// 解码Base64
-	unbased, err := base64.StdEncoding.DecodeString(imgData)
-	if err != nil {
-		return fmt.Errorf("无法解码Base64图像: %w", err)
-	}
-
-	// 解码图像
-	reader := bytes.NewReader(unbased)
-	img, _, err := image.Decode(reader)
-	if err != nil {
-		return fmt.Errorf("无法解码图像数据: %w", err)
-	}
-
-	// 创建一个可以在其上绘制的图像
-	bounds := img.Bounds()
-	rgba := image.NewRGBA(bounds)
-	draw.Draw(rgba, bounds, img, bounds.Min, draw.Src)
-
-	// 在点击/拖动位置绘制标记
-	markRadius := 10
-	x, y := int(params.Rect.X), int(params.Rect.Y)
-
-	// 绘制红色圆圈
-	for i := -markRadius; i <= markRadius; i++ {
-		for j := -markRadius; j <= markRadius; j++ {
-			if i*i+j*j <= markRadius*markRadius {
-				if x+i >= 0 && x+i < bounds.Max.X && y+j >= 0 && y+j < bounds.Max.Y {
-					rgba.Set(x+i, y+j, color.RGBA{255, 0, 0, 255})
-				}
-			}
-		}
-	}
-
-	// 保存图像
-	outFile, err := os.Create(params.OutputPath)
-	if err != nil {
-		return fmt.Errorf("无法创建输出文件: %w", err)
-	}
-	defer outFile.Close()
-
-	// 编码为PNG并保存
-	if err := png.Encode(outFile, rgba); err != nil {
-		return fmt.Errorf("无法编码和保存图像: %w", err)
-	}
-
-	return nil
-}
-
-// maskAPIKey masks the API key
-func maskAPIKey(key string) string {
-	if len(key) <= 8 {
-		return "******"
-	}
-
-	return key[:4] + "******" + key[len(key)-4:]
-}
--- a/uixt/ai/planner_gpt.go
+++ b/uixt/ai/planner_gpt.go
@@ -1,172 +0,0 @@
-package ai
-
-import (
-	"context"
-	"fmt"
-	_ "image/jpeg"
-	"strings"
-	"time"
-
-	"github.com/cloudwego/eino-ext/components/model/openai"
-	"github.com/cloudwego/eino/components/model"
-	"github.com/cloudwego/eino/schema"
-	"github.com/pkg/errors"
-	"github.com/rs/zerolog/log"
-
-	"github.com/httprunner/httprunner/v5/code"
-	"github.com/httprunner/httprunner/v5/internal/json"
-	"github.com/httprunner/httprunner/v5/uixt/types"
-)
-
-func NewPlanner(ctx context.Context) (*Planner, error) {
-	config, err := GetOpenAIModelConfig()
-	if err != nil {
-		return nil, fmt.Errorf("failed to create OpenAI config: %w", err)
-	}
-	model, err := openai.NewChatModel(ctx, config)
-	if err != nil {
-		return nil, fmt.Errorf("failed to initialize OpenAI model: %w", err)
-	}
-	return &Planner{
-		ctx:          ctx,
-		model:        model,
-		modelType:    LLMServiceTypeGPT4o,
-		systemPrompt: uiTarsPlanningPrompt, // TODO: change prompt with function calling
-	}, nil
-}
-
-type Planner struct {
-	ctx          context.Context
-	model        model.ToolCallingChatModel
-	systemPrompt string
-	modelType    LLMServiceType
-	history      ConversationHistory
-}
-
-// Call performs UI planning using Vision Language Model
-func (p *Planner) Call(opts *PlanningOptions) (*PlanningResult, error) {
-	// validate input parameters
-	if err := validatePlanningInput(opts); err != nil {
-		return nil, errors.Wrap(err, "validate planning parameters failed")
-	}
-
-	// prepare prompt
-	if len(p.history) == 0 {
-		// add system message
-		systemPrompt := uiTarsPlanningPrompt + opts.UserInstruction
-		p.history = ConversationHistory{
-			{
-				Role:    schema.System,
-				Content: systemPrompt,
-			},
-		}
-	}
-	// append user image message
-	p.history.Append(opts.Message)
-
-	// call model service, generate response
-	logRequest(p.history)
-	startTime := time.Now()
-	resp, err := p.model.Generate(p.ctx, p.history)
-	log.Info().Float64("elapsed(s)", time.Since(startTime).Seconds()).
-		Str("model", string(p.modelType)).Msg("call model service")
-	if err != nil {
-		return nil, errors.Wrap(code.LLMRequestServiceError, err.Error())
-	}
-	logResponse(resp)
-
-	// parse result
-	result, err := p.parseResult(resp, opts.Size)
-	if err != nil {
-		return nil, errors.Wrap(code.LLMParsePlanningResponseError, err.Error())
-	}
-
-	// append assistant message
-	p.history.Append(&schema.Message{
-		Role:    schema.Assistant,
-		Content: result.ActionSummary,
-	})
-
-	return result, nil
-}
-
-func (p *Planner) parseResult(msg *schema.Message, size types.Size) (*PlanningResult, error) {
-	// parse JSON format, from VLM like openai/gpt-4o
-	parseActions, jsonErr := parseJSON(msg.Content)
-	if jsonErr != nil {
-		return nil, jsonErr
-	}
-
-	// process response
-	result, err := processVLMResponse(parseActions, size)
-	if err != nil {
-		return nil, errors.Wrap(err, "process VLM response failed")
-	}
-
-	log.Info().
-		Interface("summary", result.ActionSummary).
-		Interface("actions", result.NextActions).
-		Msg("get VLM planning result")
-	return result, nil
-}
-
-// parseJSON tries to parse the response as JSON format
-func parseJSON(predictionText string) ([]ParsedAction, error) {
-	predictionText = strings.TrimSpace(predictionText)
-	if strings.HasPrefix(predictionText, "```json") && strings.HasSuffix(predictionText, "```") {
-		predictionText = strings.TrimPrefix(predictionText, "```json")
-		predictionText = strings.TrimSuffix(predictionText, "```")
-	}
-	predictionText = strings.TrimSpace(predictionText)
-
-	var response PlanningResult
-	if err := json.Unmarshal([]byte(predictionText), &response); err != nil {
-		return nil, fmt.Errorf("failed to parse VLM response: %v", err)
-	}
-
-	if response.Error != "" {
-		return nil, errors.New(response.Error)
-	}
-
-	if len(response.NextActions) == 0 {
-		return nil, errors.New("no actions returned from VLM")
-	}
-
-	// normalize actions
-	var normalizedActions []ParsedAction
-	for i := range response.NextActions {
-		// create a new variable, avoid implicit memory aliasing in for loop.
-		action := response.NextActions[i]
-		if err := normalizeAction(&action); err != nil {
-			return nil, errors.Wrap(err, "failed to normalize action")
-		}
-		normalizedActions = append(normalizedActions, action)
-	}
-
-	return normalizedActions, nil
-}
-
-// normalizeAction normalizes the coordinates in the action
-func normalizeAction(action *ParsedAction) error {
-	switch action.ActionType {
-	case "click", "drag":
-		// handle click and drag action coordinates
-		if startBox, ok := action.ActionInputs["startBox"].(string); ok {
-			normalized, err := normalizeCoordinates(startBox)
-			if err != nil {
-				return fmt.Errorf("failed to normalize startBox: %w", err)
-			}
-			action.ActionInputs["startBox"] = normalized
-		}
-
-		if endBox, ok := action.ActionInputs["endBox"].(string); ok {
-			normalized, err := normalizeCoordinates(endBox)
-			if err != nil {
-				return fmt.Errorf("failed to normalize endBox: %w", err)
-			}
-			action.ActionInputs["endBox"] = normalized
-		}
-	}
-
-	return nil
-}
--- a/uixt/ai/planner_ui_tars.go
+++ b/uixt/ai/planner_ui_tars.go
@@ -1,116 +1,38 @@
 package ai

 import (
-	"context"
 	"fmt"
 	"math"
 	"regexp"
 	"strconv"
 	"strings"
-	"time"

-	"github.com/cloudwego/eino-ext/components/model/ark"
-	"github.com/cloudwego/eino/components/model"
-	"github.com/cloudwego/eino/schema"
-	"github.com/httprunner/httprunner/v5/code"
 	"github.com/httprunner/httprunner/v5/internal/json"
 	"github.com/httprunner/httprunner/v5/uixt/types"
 	"github.com/pkg/errors"
 	"github.com/rs/zerolog/log"
 )

-func NewUITarsPlanner(ctx context.Context) (*UITarsPlanner, error) {
-	config, err := GetArkModelConfig()
-	if err != nil {
-		return nil, err
-	}
-	chatModel, err := ark.NewChatModel(ctx, config)
-	if err != nil {
-		return nil, err
-	}
-
-	return &UITarsPlanner{
-		ctx:          ctx,
-		model:        chatModel,
-		modelType:    LLMServiceTypeUITARS,
-		systemPrompt: uiTarsPlanningPrompt,
-	}, nil
+// ParsedAction represents a parsed action from the VLM response
+type ParsedAction struct {
+	ActionType   ActionType             `json:"actionType"`
+	ActionInputs map[string]interface{} `json:"actionInputs"`
+	Thought      string                 `json:"thought"`
 }

-type UITarsPlanner struct {
-	ctx          context.Context
-	model        model.ToolCallingChatModel
-	systemPrompt string
-	modelType    LLMServiceType
-	history      ConversationHistory
-}
+type ActionType string

-// Call performs UI planning using Vision Language Model
-func (p *UITarsPlanner) Call(opts *PlanningOptions) (*PlanningResult, error) {
-	// validate input parameters
-	if err := validatePlanningInput(opts); err != nil {
-		return nil, errors.Wrap(err, "validate planning parameters failed")
-	}
-
-	// prepare prompt
-	if len(p.history) == 0 {
-		// add system message
-		systemPrompt := uiTarsPlanningPrompt + opts.UserInstruction
-		p.history = ConversationHistory{
-			{
-				Role:    schema.System,
-				Content: systemPrompt,
-			},
-		}
-	}
-	// append user image message
-	p.history.Append(opts.Message)
-
-	// call model service, generate response
-	logRequest(p.history)
-	startTime := time.Now()
-	resp, err := p.model.Generate(p.ctx, p.history)
-	log.Info().Float64("elapsed(s)", time.Since(startTime).Seconds()).
-		Str("model", string(p.modelType)).Msg("call model service")
-	if err != nil {
-		return nil, errors.Wrap(code.LLMRequestServiceError, err.Error())
-	}
-	logResponse(resp)
-
-	// parse result
-	result, err := p.parseResult(resp, opts.Size)
-	if err != nil {
-		return nil, errors.Wrap(code.LLMParsePlanningResponseError, err.Error())
-	}
-
-	// append assistant message
-	p.history.Append(&schema.Message{
-		Role:    schema.Assistant,
-		Content: result.ActionSummary,
-	})
-
-	return result, nil
-}
-
-func (p *UITarsPlanner) parseResult(msg *schema.Message, size types.Size) (*PlanningResult, error) {
-	// parse Thought/Action format from UI-TARS
-	parseActions, thoughtErr := parseThoughtAction(msg.Content)
-	if thoughtErr != nil {
-		return nil, thoughtErr
-	}
-
-	// process response
-	result, err := processVLMResponse(parseActions, size)
-	if err != nil {
-		return nil, errors.Wrap(err, "process VLM response failed")
-	}
-
-	log.Info().
-		Interface("summary", result.ActionSummary).
-		Interface("actions", result.NextActions).
-		Msg("get VLM planning result")
-	return result, nil
-}
+const (
+	ActionTypeClick    ActionType = "click"
+	ActionTypeTap      ActionType = "tap"
+	ActionTypeDrag     ActionType = "drag"
+	ActionTypeSwipe    ActionType = "swipe"
+	ActionTypeWait     ActionType = "wait"
+	ActionTypeFinished ActionType = "finished"
+	ActionTypeCallUser ActionType = "call_user"
+	ActionTypeType     ActionType = "type"
+	ActionTypeScroll   ActionType = "scroll"
+)

 // parseThoughtAction parses the Thought/Action format response
 func parseThoughtAction(predictionText string) ([]ParsedAction, error) {
@@ -396,3 +318,64 @@ func validateTypeContent(action *ParsedAction) {
 		log.Warn().Msg("type action missing content parameter, set to default")
 	}
 }
+
+// parseJSON tries to parse the response as JSON format
+func parseJSON(predictionText string) ([]ParsedAction, error) {
+	predictionText = strings.TrimSpace(predictionText)
+	if strings.HasPrefix(predictionText, "```json") && strings.HasSuffix(predictionText, "```") {
+		predictionText = strings.TrimPrefix(predictionText, "```json")
+		predictionText = strings.TrimSuffix(predictionText, "```")
+	}
+	predictionText = strings.TrimSpace(predictionText)
+
+	var response PlanningResult
+	if err := json.Unmarshal([]byte(predictionText), &response); err != nil {
+		return nil, fmt.Errorf("failed to parse VLM response: %v", err)
+	}
+
+	if response.Error != "" {
+		return nil, errors.New(response.Error)
+	}
+
+	if len(response.NextActions) == 0 {
+		return nil, errors.New("no actions returned from VLM")
+	}
+
+	// normalize actions
+	var normalizedActions []ParsedAction
+	for i := range response.NextActions {
+		// create a new variable, avoid implicit memory aliasing in for loop.
+		action := response.NextActions[i]
+		if err := normalizeAction(&action); err != nil {
+			return nil, errors.Wrap(err, "failed to normalize action")
+		}
+		normalizedActions = append(normalizedActions, action)
+	}
+
+	return normalizedActions, nil
+}
+
+// normalizeAction normalizes the coordinates in the action
+func normalizeAction(action *ParsedAction) error {
+	switch action.ActionType {
+	case "click", "drag":
+		// handle click and drag action coordinates
+		if startBox, ok := action.ActionInputs["startBox"].(string); ok {
+			normalized, err := normalizeCoordinates(startBox)
+			if err != nil {
+				return fmt.Errorf("failed to normalize startBox: %w", err)
+			}
+			action.ActionInputs["startBox"] = normalized
+		}
+
+		if endBox, ok := action.ActionInputs["endBox"].(string); ok {
+			normalized, err := normalizeCoordinates(endBox)
+			if err != nil {
+				return fmt.Errorf("failed to normalize endBox: %w", err)
+			}
+			action.ActionInputs["endBox"] = normalized
+		}
+	}
+
+	return nil
+}
--- a/uixt/ai/planner_prompts.go
+++ b/uixt/ai/planner_prompts.go
@@ -27,3 +27,5 @@ finished(content='xxx') # Use escape characters \\', \\", and \\n in content par

 ## User Instruction
 `
+
+const defaultPlanningResponseJsonFormat = ``
--- a/uixt/ai/planner_test.go
+++ b/uixt/ai/planner_test.go
@@ -1,25 +1,20 @@
 package ai

 import (
-	"bytes"
 	"context"
-	"encoding/base64"
-	"fmt"
-	"image"
-	"image/jpeg"
-	"image/png"
-	"os"
 	"testing"

 	"github.com/cloudwego/eino/schema"
 	"github.com/httprunner/httprunner/v5/code"
+	"github.com/httprunner/httprunner/v5/internal/builtin"
+	"github.com/httprunner/httprunner/v5/uixt/option"
 	"github.com/httprunner/httprunner/v5/uixt/types"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 )

 func TestVLMPlanning(t *testing.T) {
-	imageBase64, size, err := loadImage("testdata/llk_1.png")
+	imageBase64, size, err := builtin.LoadImage("testdata/llk_1.png")
 	require.NoError(t, err)

 	userInstruction := `连连看是一款经典的益智消除类小游戏，通常以图案或图标为主要元素。以下是连连看的基本规则说明：
@@ -35,7 +30,10 @@ func TestVLMPlanning(t *testing.T) {

 	userInstruction += "\n\n请基于以上游戏规则，给出下一步可点击的两个图标坐标"

-	planner, err := NewUITarsPlanner(context.Background())
+	modelConfig, err := GetModelConfig(option.LLMServiceTypeUITARS)
+	require.NoError(t, err)
+
+	planner, err := NewPlanner(context.Background(), modelConfig)
 	require.NoError(t, err)

 	opts := &PlanningOptions{
@@ -100,12 +98,15 @@ func TestVLMPlanning(t *testing.T) {
 }

 func TestXHSPlanning(t *testing.T) {
-	imageBase64, size, err := loadImage("testdata/xhs-feed.jpeg")
+	imageBase64, size, err := builtin.LoadImage("testdata/xhs-feed.jpeg")
 	require.NoError(t, err)

 	userInstruction := "点击第二个帖子的作者头像"

-	planner, err := NewUITarsPlanner(context.Background())
+	modelConfig, err := GetModelConfig(option.LLMServiceTypeUITARS)
+	require.NoError(t, err)
+
+	planner, err := NewPlanner(context.Background(), modelConfig)
 	require.NoError(t, err)

 	opts := &PlanningOptions{
@@ -170,12 +171,15 @@ func TestXHSPlanning(t *testing.T) {
 }

 func TestChatList(t *testing.T) {
-	imageBase64, size, err := loadImage("testdata/chat_list.jpeg")
+	imageBase64, size, err := builtin.LoadImage("testdata/chat_list.jpeg")
 	require.NoError(t, err)

 	userInstruction := "请结合图片的文字信息，请告诉我一共有多少个群聊，哪些群聊右下角有绿点"

-	planner, err := NewUITarsPlanner(context.Background())
+	modelConfig, err := GetModelConfig(option.LLMServiceTypeUITARS)
+	require.NoError(t, err)
+
+	planner, err := NewPlanner(context.Background(), modelConfig)
 	require.NoError(t, err)

 	opts := &PlanningOptions{
@@ -206,7 +210,10 @@ func TestHandleSwitch(t *testing.T) {
 	userInstruction := "发送框下方的联网搜索开关是开启状态" // 点击开启联网搜索开关
 	// 检查发送框下方的联网搜索开关，蓝色为开启状态，灰色为关闭状态；若开关处于关闭状态，则点击进行开启

-	planner, err := NewUITarsPlanner(context.Background())
+	modelConfig, err := GetModelConfig(option.LLMServiceTypeUITARS)
+	require.NoError(t, err)
+
+	planner, err := NewPlanner(context.Background(), modelConfig)
 	require.NoError(t, err)

 	testCases := []struct {
@@ -219,7 +226,7 @@ func TestHandleSwitch(t *testing.T) {
 	}

 	for _, tc := range testCases {
-		imageBase64, size, err := loadImage(tc.imageFile)
+		imageBase64, size, err := builtin.LoadImage(tc.imageFile)
 		require.NoError(t, err)

 		opts := &PlanningOptions{
@@ -250,7 +257,7 @@ func TestHandleSwitch(t *testing.T) {
 }

 func TestValidateInput(t *testing.T) {
-	imageBase64, size, err := loadImage("testdata/popup_risk_warning.png")
+	imageBase64, size, err := builtin.LoadImage("testdata/popup_risk_warning.png")
 	require.NoError(t, err)

 	tests := []struct {
@@ -375,87 +382,18 @@ func TestProcessVLMResponse(t *testing.T) {
 	}
 }

-func TestSavePositionImg(t *testing.T) {
-	imageBase64, _, err := loadImage("testdata/popup_risk_warning.png")
-	require.NoError(t, err)
-
-	params := struct {
-		InputImgBase64 string
-		Rect           struct {
-			X float64
-			Y float64
-		}
-		OutputPath string
-	}{
-		InputImgBase64: imageBase64,
-		Rect: struct {
-			X float64
-			Y float64
-		}{
-			X: 100,
-			Y: 100,
-		},
-		OutputPath: "testdata/output.png",
-	}
-
-	err = SavePositionImg(params)
-	require.NoError(t, err)
-
-	// cleanup
-	defer os.Remove(params.OutputPath)
-}
-
 func TestLoadImage(t *testing.T) {
 	// Test PNG image
-	pngBase64, pngSize, err := loadImage("testdata/llk_1.png")
+	pngBase64, pngSize, err := builtin.LoadImage("testdata/llk_1.png")
 	require.NoError(t, err)
 	assert.NotEmpty(t, pngBase64)
 	assert.Greater(t, pngSize.Width, 0)
 	assert.Greater(t, pngSize.Height, 0)

 	// Test JPEG image
-	jpegBase64, jpegSize, err := loadImage("testdata/xhs-feed.jpeg")
+	jpegBase64, jpegSize, err := builtin.LoadImage("testdata/xhs-feed.jpeg")
 	require.NoError(t, err)
 	assert.NotEmpty(t, jpegBase64)
 	assert.Greater(t, jpegSize.Width, 0)
 	assert.Greater(t, jpegSize.Height, 0)
 }
-
-// loadImage loads image and returns base64 encoded string
-func loadImage(imagePath string) (base64Str string, size types.Size, err error) {
-	// Read the image file
-	imageFile, err := os.OpenFile(imagePath, os.O_RDONLY, 0o600)
-	if err != nil {
-		return "", types.Size{}, fmt.Errorf("failed to open image file: %w", err)
-	}
-	defer imageFile.Close()
-
-	// Decode the image to get its resolution
-	imageData, format, err := image.Decode(imageFile)
-	if err != nil {
-		return "", types.Size{}, fmt.Errorf("failed to decode image: %w", err)
-	}
-
-	// Get the resolution of the image
-	width := imageData.Bounds().Dx()
-	height := imageData.Bounds().Dy()
-	size = types.Size{Width: width, Height: height}
-
-	// Convert image to base64
-	buf := new(bytes.Buffer)
-	// 根据图像格式选择正确的编码器
-	if format == "jpeg" || format == "jpg" {
-		if err := jpeg.Encode(buf, imageData, nil); err != nil {
-			return "", types.Size{}, fmt.Errorf("failed to encode image to buffer: %w", err)
-		}
-	} else {
-		// 默认使用 PNG 编码
-		if err := png.Encode(buf, imageData); err != nil {
-			return "", types.Size{}, fmt.Errorf("failed to encode image to buffer: %w", err)
-		}
-	}
-	base64Str = fmt.Sprintf("data:image/%s;base64,%s", format,
-		base64.StdEncoding.EncodeToString(buf.Bytes()))
-
-	return base64Str, size, nil
-}
--- a/uixt/android_driver_adb.go
+++ b/uixt/android_driver_adb.go
@@ -307,14 +307,16 @@ func (ad *ADBDriver) TapXY(x, y float64, opts ...option.ActionOption) error {

 func (ad *ADBDriver) TapAbsXY(x, y float64, opts ...option.ActionOption) error {
 	log.Info().Float64("x", x).Float64("y", y).Msg("ADBDriver.TapAbsXY")
-	actionOptions := option.NewActionOptions(opts...)
-	x, y = actionOptions.ApplyTapOffset(x, y)
+	var err error
+	x, y, err = handlerTapAbsXY(ad, x, y, opts...)
+	if err != nil {
+		return err
+	}

 	// adb shell input tap x y
 	xStr := fmt.Sprintf("%.1f", x)
 	yStr := fmt.Sprintf("%.1f", y)
-	_, err := ad.runShellCommand(
-		"input", "tap", xStr, yStr)
+	_, err = ad.runShellCommand("input", "tap", xStr, yStr)
 	if err != nil {
 		return errors.Wrap(err, fmt.Sprintf("tap <%s, %s> failed", xStr, yStr))
 	}
@@ -324,12 +326,10 @@ func (ad *ADBDriver) TapAbsXY(x, y float64, opts ...option.ActionOption) error {
 func (ad *ADBDriver) DoubleTap(x, y float64, opts ...option.ActionOption) error {
 	log.Info().Float64("x", x).Float64("y", y).Msg("ADBDriver.DoubleTap")
 	var err error
-	x, y, err = convertToAbsolutePoint(ad, x, y)
+	x, y, err = handlerDoubleTap(ad, x, y, opts...)
 	if err != nil {
 		return err
 	}
-	actionOptions := option.NewActionOptions(opts...)
-	x, y = actionOptions.ApplyTapOffset(x, y)

 	// adb shell input tap x y
 	xStr := fmt.Sprintf("%.1f", x)
@@ -372,13 +372,13 @@ func (ad *ADBDriver) TouchAndHold(x, y float64, opts ...option.ActionOption) (er
 func (ad *ADBDriver) Drag(fromX, fromY, toX, toY float64, opts ...option.ActionOption) (err error) {
 	log.Info().Float64("fromX", fromX).Float64("fromY", fromY).
 		Float64("toX", toX).Float64("toY", toY).Msg("ADBDriver.Drag")
-	actionOptions := option.NewActionOptions(opts...)
-	fromX, fromY, toX, toY, err = convertToAbsoluteCoordinates(ad, fromX, fromY, toX, toY)
+
+	fromX, fromY, toX, toY, err = handlerDrag(ad, fromX, fromY, toX, toY, opts...)
 	if err != nil {
 		return err
 	}
-	fromX, fromY, toX, toY = actionOptions.ApplySwipeOffset(fromX, fromY, toX, toY)

+	actionOptions := option.NewActionOptions(opts...)
 	duration := 200.0
 	if actionOptions.Duration > 0 {
 		duration = actionOptions.Duration * 1000
@@ -404,12 +404,10 @@ func (ad *ADBDriver) Swipe(fromX, fromY, toX, toY float64, opts ...option.Action
 	log.Info().Float64("fromX", fromX).Float64("fromY", fromY).
 		Float64("toX", toX).Float64("toY", toY).Msg("ADBDriver.Swipe")
 	var err error
-	fromX, fromY, toX, toY, err = convertToAbsoluteCoordinates(ad, fromX, fromY, toX, toY)
+	fromX, fromY, toX, toY, err = handlerSwipe(ad, fromX, fromY, toX, toY)
 	if err != nil {
 		return err
 	}
-	actionOptions := option.NewActionOptions(opts...)
-	fromX, fromY, toX, toY = actionOptions.ApplySwipeOffset(fromX, fromY, toX, toY)

 	// adb shell input swipe fromX fromY toX toY
 	_, err = ad.runShellCommand(
@@ -701,7 +699,7 @@ func (ad *ADBDriver) StopCaptureLog() (result interface{}, err error) {
 			return pointRes, nil
 		}

-		reader, err := os.OpenFile(files[0], os.O_RDONLY, 0o600)
+		reader, err := os.Open(files[0])
 		if err != nil {
 			log.Info().Msg("open File error")
 			return pointRes, nil
--- a/uixt/android_driver_uia2.go
+++ b/uixt/android_driver_uia2.go
@@ -256,12 +256,10 @@ func (ud *UIA2Driver) Orientation() (orientation types.Orientation, err error) {
 func (ud *UIA2Driver) DoubleTap(x, y float64, opts ...option.ActionOption) error {
 	log.Info().Float64("x", x).Float64("y", y).Msg("UIA2Driver.DoubleTap")
 	var err error
-	x, y, err = convertToAbsolutePoint(ud, x, y)
+	x, y, err = handlerDoubleTap(ud, x, y, opts...)
 	if err != nil {
 		return err
 	}
-	actionOptions := option.NewActionOptions(opts...)
-	x, y = actionOptions.ApplyTapOffset(x, y)

 	data := map[string]interface{}{
 		"actions": []interface{}{
@@ -298,13 +296,19 @@ func (ud *UIA2Driver) TapXY(x, y float64, opts ...option.ActionOption) error {
 func (ud *UIA2Driver) TapAbsXY(x, y float64, opts ...option.ActionOption) error {
 	log.Info().Float64("x", x).Float64("y", y).Msg("UIA2Driver.TapAbsXY")
 	// register(postHandler, new Tap("/wd/hub/session/:sessionId/appium/tap"))
-	actionOptions := option.NewActionOptions(opts...)
-	x, y = actionOptions.ApplyTapOffset(x, y)

+	var err error
+	x, y, err = handlerTapAbsXY(ud, x, y, opts...)
+	if err != nil {
+		return err
+	}
+
+	actionOptions := option.NewActionOptions(opts...)
 	duration := 100.0
 	if actionOptions.PressDuration > 0 {
 		duration = actionOptions.PressDuration * 1000 // convert to ms
 	}
+
 	data := map[string]interface{}{
 		"actions": []interface{}{
 			map[string]interface{}{
@@ -323,7 +327,7 @@ func (ud *UIA2Driver) TapAbsXY(x, y float64, opts ...option.ActionOption) error
 	option.MergeOptions(data, opts...)

 	urlStr := fmt.Sprintf("/session/%s/actions/tap", ud.Session.ID)
-	_, err := ud.Session.POST(data, urlStr)
+	_, err = ud.Session.POST(data, urlStr)
 	return err
 }

@@ -355,13 +359,12 @@ func (ud *UIA2Driver) TouchAndHold(x, y float64, opts ...option.ActionOption) (e
 func (ud *UIA2Driver) Drag(fromX, fromY, toX, toY float64, opts ...option.ActionOption) error {
 	log.Info().Float64("fromX", fromX).Float64("fromY", fromY).
 		Float64("toX", toX).Float64("toY", toY).Msg("UIA2Driver.Drag")
+
 	var err error
-	fromX, fromY, toX, toY, err = convertToAbsoluteCoordinates(ud, fromX, fromY, toX, toY)
+	fromX, fromY, toX, toY, err = handlerDrag(ud, fromX, fromY, toX, toY, opts...)
 	if err != nil {
 		return err
 	}
-	actionOptions := option.NewActionOptions(opts...)
-	fromX, fromY, toX, toY = actionOptions.ApplySwipeOffset(fromX, fromY, toX, toY)

 	data := map[string]interface{}{
 		"startX": fromX,
@@ -387,13 +390,11 @@ func (ud *UIA2Driver) Swipe(fromX, fromY, toX, toY float64, opts ...option.Actio
 	log.Info().Float64("fromX", fromX).Float64("fromY", fromY).
 		Float64("toX", toX).Float64("toY", toY).Msg("UIA2Driver.Swipe")
 	var err error
-	actionOptions := option.NewActionOptions(opts...)
-	fromX, fromY, toX, toY, err = convertToAbsoluteCoordinates(ud, fromX, fromY, toX, toY)
+	fromX, fromY, toX, toY, err = handlerSwipe(ud, fromX, fromY, toX, toY)
 	if err != nil {
 		return err
 	}
-	fromX, fromY, toX, toY = actionOptions.ApplySwipeOffset(fromX, fromY, toX, toY)
-
+	actionOptions := option.NewActionOptions(opts...)
 	duration := 200.0
 	if actionOptions.PressDuration > 0 {
 		duration = actionOptions.PressDuration * 1000 // ms
--- a/uixt/android_test.go
+++ b/uixt/android_test.go
@@ -11,7 +11,6 @@ import (
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"

-	"github.com/httprunner/httprunner/v5/uixt/ai"
 	"github.com/httprunner/httprunner/v5/uixt/option"
 	"github.com/httprunner/httprunner/v5/uixt/types"
 )
@@ -23,10 +22,12 @@ func setupADBDriverExt(t *testing.T) *XTDriver {
 	device.Options.LogOn = false
 	driver, err := device.NewDriver()
 	require.Nil(t, err)
-	return NewXTDriver(driver,
-		ai.WithCVService(ai.CVServiceTypeVEDEM),
-		ai.WithLLMService(ai.LLMServiceTypeUITARS),
+	driverExt, err := NewXTDriver(driver,
+		option.WithCVService(option.CVServiceTypeVEDEM),
+		option.WithLLMService(option.LLMServiceTypeUITARS),
 	)
+	require.Nil(t, err)
+	return driverExt
 }

 func setupUIA2DriverExt(t *testing.T) *XTDriver {
@@ -36,8 +37,10 @@ func setupUIA2DriverExt(t *testing.T) *XTDriver {
 	device.Options.LogOn = false
 	driver, err := device.NewDriver()
 	require.Nil(t, err)
-	return NewXTDriver(driver,
-		ai.WithCVService(ai.CVServiceTypeVEDEM))
+	driverExt, err := NewXTDriver(driver,
+		option.WithCVService(option.CVServiceTypeVEDEM))
+	require.Nil(t, err)
+	return driverExt
 }

 func TestDevice_Android_GetPackageInfo(t *testing.T) {
--- a/uixt/browser_driver.go
+++ b/uixt/browser_driver.go
@@ -103,7 +103,13 @@ func NewBrowserDriver(device *BrowserDevice) (driver *BrowserDriver, err error)
 	return driver, nil
 }

-func (wd *BrowserDriver) Drag(fromX, fromY, toX, toY float64, options ...option.ActionOption) (err error) {
+func (wd *BrowserDriver) Drag(fromX, fromY, toX, toY float64, opts ...option.ActionOption) error {
+	var err error
+	fromX, fromY, toX, toY, err = handlerDrag(wd, fromX, fromY, toX, toY, opts...)
+	if err != nil {
+		return err
+	}
+
 	data := map[string]interface{}{
 		"from_x": fromX,
 		"from_y": fromY,
@@ -111,14 +117,13 @@ func (wd *BrowserDriver) Drag(fromX, fromY, toX, toY float64, options ...option.
 		"to_y":   toY,
 	}

-	actionOptions := option.NewActionOptions(options...)
-
+	actionOptions := option.NewActionOptions(opts...)
 	if actionOptions.Duration > 0 {
 		data["duration"] = actionOptions.Duration
 	}

 	_, err = wd.HttpPOST(data, wd.sessionId, "ui/drag")
-	return
+	return err
 }

 func (wd *BrowserDriver) AppLaunch(packageName string) (err error) {
@@ -517,28 +522,40 @@ func (wd *BrowserDriver) Tap(x, y float64, options ...option.ActionOption) error
 	return wd.TapFloat(x, y, options...)
 }

-func (wd *BrowserDriver) TapFloat(x, y float64, options ...option.ActionOption) error {
-	actionOptions := option.NewActionOptions(options...)
+func (wd *BrowserDriver) TapFloat(x, y float64, opts ...option.ActionOption) error {
+	var err error
+	x, y, err = handlerTapAbsXY(wd, x, y, opts...)
+	if err != nil {
+		return err
+	}
+
+	actionOptions := option.NewActionOptions(opts...)
 	duration := 0.1
 	if actionOptions.Duration > 0 {
 		duration = actionOptions.Duration
 	}
+
 	data := map[string]interface{}{
 		"x":        x,
 		"y":        y,
 		"duration": duration,
 	}
-	_, err := wd.HttpPOST(data, wd.sessionId, "ui/tap")
+	_, err = wd.HttpPOST(data, wd.sessionId, "ui/tap")
 	return err
 }

 // DoubleTap Sends a double tap event at the coordinate.
 func (wd *BrowserDriver) DoubleTap(x, y float64, options ...option.ActionOption) error {
+	var err error
+	x, y, err = handlerDoubleTap(wd, x, y, options...)
+	if err != nil {
+		return err
+	}
 	data := map[string]interface{}{
 		"x": x,
 		"y": y,
 	}
-	_, err := wd.HttpPOST(data, wd.sessionId, "ui/double_tap")
+	_, err = wd.HttpPOST(data, wd.sessionId, "ui/double_tap")
 	return err
 }

--- a/uixt/demo/main_test.go
+++ b/uixt/demo/main_test.go
@@ -9,7 +9,6 @@ import (
 	"github.com/stretchr/testify/assert"

 	"github.com/httprunner/httprunner/v5/uixt"
-	"github.com/httprunner/httprunner/v5/uixt/ai"
 	"github.com/httprunner/httprunner/v5/uixt/option"
 )

@@ -23,9 +22,10 @@ func TestIOSDemo(t *testing.T) {

 	driver, err := device.NewDriver()
 	assert.Nil(t, err)
-	driverExt := uixt.NewXTDriver(driver,
-		ai.WithCVService(ai.CVServiceTypeVEDEM),
+	driverExt, err := uixt.NewXTDriver(driver,
+		option.WithCVService(option.CVServiceTypeVEDEM),
 	)
+	assert.Nil(t, err)

 	// release session
 	defer func() {
--- a/uixt/driver.go
+++ b/uixt/driver.go
@@ -8,6 +8,7 @@ import (
 	"github.com/httprunner/httprunner/v5/uixt/ai"
 	"github.com/httprunner/httprunner/v5/uixt/option"
 	"github.com/httprunner/httprunner/v5/uixt/types"
+	"github.com/rs/zerolog/log"
 )

 var (
@@ -81,16 +82,30 @@ type IDriver interface {
 	StopCaptureLog() (result interface{}, err error)
 }

-func NewXTDriver(driver IDriver, opts ...ai.AIServiceOption) *XTDriver {
-	services := ai.NewAIService(opts...)
+func NewXTDriver(driver IDriver, opts ...option.AIServiceOption) (*XTDriver, error) {
 	driverExt := &XTDriver{
-		IDriver:    driver,
-		CVService:  services.ICVService,
-		LLMService: services.ILLMService,
-
-		screenResults: make([]*ScreenResult, 0),
+		IDriver: driver,
 	}
-	return driverExt
+
+	services := option.NewAIServiceOptions(opts...)
+
+	var err error
+	if services.CVService != "" {
+		driverExt.CVService, err = ai.NewCVService(services.CVService)
+		if err != nil {
+			log.Error().Err(err).Msg("init vedem image service failed")
+			return nil, err
+		}
+	}
+	if services.LLMService != "" {
+		driverExt.LLMService, err = ai.NewLLMService(services.LLMService)
+		if err != nil {
+			log.Error().Err(err).Msg("init llm service failed")
+			return nil, err
+		}
+	}
+
+	return driverExt, nil
 }

 // XTDriver = IDriver + AI
@@ -98,7 +113,4 @@ type XTDriver struct {
 	IDriver
 	CVService  ai.ICVService  // OCR/CV
 	LLMService ai.ILLMService // LLM
-
-	// cache screenshot results
-	screenResults []*ScreenResult
 }
--- a/uixt/driver_action.go
+++ b/uixt/driver_action.go
@@ -42,6 +42,7 @@ const (
 	ACTION_TapByCV     ActionMethod = "tap_cv"
 	ACTION_DoubleTapXY ActionMethod = "double_tap_xy"
 	ACTION_Swipe       ActionMethod = "swipe"
+	ACTION_Drag        ActionMethod = "drag"
 	ACTION_Input       ActionMethod = "input"
 	ACTION_Back        ActionMethod = "back"
 	ACTION_KeyCode     ActionMethod = "keycode"
--- a/uixt/driver_ext_screenshot.go
+++ b/uixt/driver_ext_screenshot.go
@@ -4,12 +4,16 @@ import (
 	"bytes"
 	"fmt"
 	"image"
+	"image/color"
+	"image/draw"
 	"image/gif"
 	"image/jpeg"
 	"image/png"
+	"math"
 	"os"
 	"path/filepath"
 	"strings"
+	"time"

 	"github.com/pkg/errors"
 	"github.com/rs/zerolog/log"
@@ -129,8 +133,9 @@ func (dExt *XTDriver) GetScreenResult(opts ...option.ActionOption) (screenResult
 		}
 	}

-	// cache screen result
-	dExt.screenResults = append(dExt.screenResults, screenResult)
+	// save screen result to session
+	session := dExt.GetSession()
+	session.screenResults = append(session.screenResults, screenResult)

 	log.Debug().
 		Str("imagePath", imagePath).
@@ -294,3 +299,255 @@ func compressImageBuffer(raw *bytes.Buffer) (compressed *bytes.Buffer, err error
 	// return compressed image buffer
 	return &buf, nil
 }
+
+// MarkUIOperation add operation mark for UI operation
+func MarkUIOperation(driver IDriver, actionType ActionMethod, actionCoordinates []float64) error {
+	if actionType == "" || len(actionCoordinates) == 0 {
+		return nil
+	}
+	start := time.Now()
+
+	// get screenshot
+	compressedBufSource, err := driver.ScreenShot()
+	if err != nil {
+		return err
+	}
+
+	// create screenshot save path
+	timestamp := builtin.GenNameWithTimestamp("action_%d")
+	var imagePath string
+
+	if actionType == ACTION_TapAbsXY || actionType == ACTION_DoubleTapXY {
+		if len(actionCoordinates) != 2 {
+			return fmt.Errorf("invalid tap action coordinates: %v", actionCoordinates)
+		}
+		imagePath = filepath.Join(
+			config.GetConfig().ScreenShotsPath,
+			fmt.Sprintf("%s_%s.png", timestamp, actionType),
+		)
+		x, y := actionCoordinates[0], actionCoordinates[1]
+		point := image.Point{X: int(x), Y: int(y)}
+		err = SaveImageWithCircleMarker(compressedBufSource, point, imagePath)
+	} else if actionType == ACTION_Swipe || actionType == ACTION_Drag {
+		if len(actionCoordinates) != 4 {
+			return fmt.Errorf("invalid swipe action coordinates: %v", actionCoordinates)
+		}
+		imagePath = filepath.Join(
+			config.GetConfig().ScreenShotsPath,
+			fmt.Sprintf("%s_%s.png", timestamp, actionType),
+		)
+		fromX, fromY := actionCoordinates[0], actionCoordinates[1]
+		toX, toY := actionCoordinates[2], actionCoordinates[3]
+		from := image.Point{X: int(fromX), Y: int(fromY)}
+		to := image.Point{X: int(toX), Y: int(toY)}
+		err = SaveImageWithArrowMarker(compressedBufSource, from, to, imagePath)
+	}
+	if err != nil {
+		log.Error().Err(err).
+			Int64("duration(ms)", time.Since(start).Milliseconds()).
+			Msg("mark UI operation failed")
+		return err
+	}
+
+	if imagePath != "" {
+		log.Info().Str("operation", string(actionType)).
+			Str("imagePath", imagePath).
+			Int64("duration(ms)", time.Since(start).Milliseconds()).
+			Msg("mark UI operation success")
+
+		// save screenshot to session
+		session := driver.GetSession()
+		session.screenResults = append(session.screenResults, &ScreenResult{
+			bufSource: compressedBufSource,
+			ImagePath: imagePath,
+		})
+	}
+
+	return nil
+}
+
+// SaveImageWithCircleMarker saves an image with circle marker
+func SaveImageWithCircleMarker(imgBuf *bytes.Buffer, point image.Point, outputPath string) error {
+	img, _, err := image.Decode(imgBuf)
+	if err != nil {
+		return fmt.Errorf("failed to decode image data: %w", err)
+	}
+	bounds := img.Bounds()
+	rgba := image.NewRGBA(bounds)
+	draw.Draw(rgba, bounds, img, bounds.Min, draw.Src)
+
+	// draw a red circle at the tap point
+	centerX := point.X
+	centerY := point.Y
+	radius := 20
+	lineWidth := 5
+	red := color.RGBA{255, 0, 0, 255}
+
+	for angle := 0.0; angle < 2*math.Pi; angle += 0.01 {
+		for w := 0; w < lineWidth; w++ {
+			r := float64(radius - w)
+			x := int(float64(centerX) + r*math.Cos(angle))
+			y := int(float64(centerY) + r*math.Sin(angle))
+			if x >= 0 && x < bounds.Max.X && y >= 0 && y < bounds.Max.Y {
+				rgba.Set(x, y, red)
+			}
+		}
+	}
+
+	outFile, err := os.Create(outputPath)
+	if err != nil {
+		return fmt.Errorf("failed to create output file: %w", err)
+	}
+	defer outFile.Close()
+	if err := png.Encode(outFile, rgba); err != nil {
+		return fmt.Errorf("failed to encode and save image: %w", err)
+	}
+	return nil
+}
+
+// SaveImageWithArrowMarker saves an image with an arrow marker
+func SaveImageWithArrowMarker(imgBuf *bytes.Buffer, from, to image.Point, outputPath string) error {
+	img, _, err := image.Decode(imgBuf)
+	if err != nil {
+		return fmt.Errorf("failed to decode image data: %w", err)
+	}
+	bounds := img.Bounds()
+	rgba := image.NewRGBA(bounds)
+	draw.Draw(rgba, bounds, img, bounds.Min, draw.Src)
+	drawArrow(rgba, from, to, color.RGBA{255, 0, 0, 255}, 5)
+	outFile, err := os.Create(outputPath)
+	if err != nil {
+		return fmt.Errorf("failed to create output file: %w", err)
+	}
+	defer outFile.Close()
+	if err := png.Encode(outFile, rgba); err != nil {
+		return fmt.Errorf("failed to encode and save image: %w", err)
+	}
+	return nil
+}
+
+// drawArrow draws an arrow from 'from' to 'to' on the image
+func drawArrow(rgba *image.RGBA, from, to image.Point, color color.RGBA, lineWidth int) {
+	bounds := rgba.Bounds()
+	dx, dy := to.X-from.X, to.Y-from.Y
+	steps := int(math.Sqrt(float64(dx*dx + dy*dy)))
+	if steps == 0 {
+		steps = 1
+	}
+	stepX, stepY := float64(dx)/float64(steps), float64(dy)/float64(steps)
+	// main line
+	for i := 0; i < steps; i++ {
+		x := int(float64(from.X) + stepX*float64(i))
+		y := int(float64(from.Y) + stepY*float64(i))
+		for w := 0; w < lineWidth; w++ {
+			offsetX, offsetY := 0, 0
+			if math.Abs(stepX) > math.Abs(stepY) {
+				offsetY = w - lineWidth/2
+			} else {
+				offsetX = w - lineWidth/2
+			}
+			drawX, drawY := x+offsetX, y+offsetY
+			if drawX >= 0 && drawX < bounds.Max.X && drawY >= 0 && drawY < bounds.Max.Y {
+				rgba.Set(drawX, drawY, color)
+			}
+		}
+	}
+	// arrow head
+	arrowLength := float64(steps) * 0.15
+	if arrowLength < 10 {
+		arrowLength = 10
+	} else if arrowLength > 30 {
+		arrowLength = 30
+	}
+	head := calculateArrowHead(float64(from.X), float64(from.Y), float64(to.X), float64(to.Y), arrowLength)
+	if head != nil {
+		for _, point := range head[:2] {
+			drawLineInImage(rgba, to.X, to.Y, int(point.X), int(point.Y), color, lineWidth, bounds)
+		}
+		for _, point := range head[1:] {
+			drawLineInImage(rgba, to.X, to.Y, int(point.X), int(point.Y), color, lineWidth, bounds)
+		}
+	}
+}
+
+// calculateArrowHead calculates the endpoint and arrowhead coordinates
+func calculateArrowHead(fromX, fromY, toX, toY float64, arrowLength float64) []struct{ X, Y float64 } {
+	// calculate direction vector
+	dx, dy := toX-fromX, toY-fromY
+	// calculate distance
+	length := math.Sqrt(dx*dx + dy*dy)
+	if length < 1e-6 {
+		return nil
+	}
+
+	// unit vector
+	dx, dy = dx/length, dy/length
+
+	// calculate orthogonal vector of arrow direction (counterclockwise 90 degrees)
+	orthX, orthY := -dy, dx
+
+	// calculate two wing points of arrow
+	headWidth := arrowLength * 0.5
+	backX, backY := toX-dx*arrowLength, toY-dy*arrowLength
+
+	// two wing points of arrow
+	leftWingX, leftWingY := backX+orthX*headWidth, backY+orthY*headWidth
+	rightWingX, rightWingY := backX-orthX*headWidth, backY-orthY*headWidth
+
+	return []struct{ X, Y float64 }{
+		{leftWingX, leftWingY},
+		{toX, toY},
+		{rightWingX, rightWingY},
+	}
+}
+
+// drawLineInImage draws a line on the image
+func drawLineInImage(img *image.RGBA, x0, y0, x1, y1 int, lineColor color.RGBA, lineWidth int, bounds image.Rectangle) {
+	// use Bresenham algorithm to draw line
+	dx, dy := math.Abs(float64(x1-x0)), math.Abs(float64(y1-y0))
+	sx, sy := 1, 1
+	if x0 >= x1 {
+		sx = -1
+	}
+	if y0 >= y1 {
+		sy = -1
+	}
+	err := dx - dy
+
+	for {
+		// draw point (consider line width)
+		for w := 0; w < lineWidth; w++ {
+			offsetX, offsetY := 0, 0
+
+			// decide offset direction based on line angle
+			if dx > dy {
+				// more horizontal line
+				offsetY = w - lineWidth/2
+			} else {
+				// more vertical line
+				offsetX = w - lineWidth/2
+			}
+
+			drawX, drawY := x0+offsetX, y0+offsetY
+			if drawX >= 0 && drawX < bounds.Max.X && drawY >= 0 && drawY < bounds.Max.Y {
+				img.Set(drawX, drawY, lineColor)
+			}
+		}
+
+		// end of line
+		if x0 == x1 && y0 == y1 {
+			break
+		}
+
+		// calculate next point
+		e2 := 2 * err
+		if e2 > -dy {
+			err = err - dy
+			x0 = x0 + sx
+		}
+		if e2 < dx {
+			err = err + dx
+			y0 = y0 + sy
+		}
+	}
+}
--- a/uixt/driver_ext_test.go
+++ b/uixt/driver_ext_test.go
@@ -3,10 +3,12 @@
 package uixt

 import (
+	"bytes"
+	"image"
+	"os"
 	"testing"
 	"time"

-	"github.com/httprunner/httprunner/v5/uixt/ai"
 	"github.com/httprunner/httprunner/v5/uixt/option"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
@@ -17,8 +19,9 @@ func TestDriverExt_NewMethod1(t *testing.T) {
 	require.Nil(t, err)
 	driver, err := device.NewDriver()
 	require.Nil(t, err)
-	driverExt := NewXTDriver(driver,
-		ai.WithCVService(ai.CVServiceTypeVEDEM))
+	driverExt, err := NewXTDriver(driver,
+		option.WithCVService(option.CVServiceTypeVEDEM))
+	require.Nil(t, err)
 	driverExt.TapByOCR("推荐")
 }

@@ -27,16 +30,18 @@ func TestDriverExt_NewMethod2(t *testing.T) {
 	require.Nil(t, err)
 	driver, err := NewUIA2Driver(device)
 	require.Nil(t, err)
-	driverExt := NewXTDriver(driver,
-		ai.WithCVService(ai.CVServiceTypeVEDEM))
+	driverExt, err := NewXTDriver(driver,
+		option.WithCVService(option.CVServiceTypeVEDEM))
+	require.Nil(t, err)
 	driverExt.TapByOCR("推荐")
 }

 func TestDriverExt(t *testing.T) {
 	device, _ := NewAndroidDevice()
 	driver, _ := NewADBDriver(device)
-	driverExt := NewXTDriver(driver,
-		ai.WithCVService(ai.CVServiceTypeVEDEM))
+	driverExt, err := NewXTDriver(driver,
+		option.WithCVService(option.CVServiceTypeVEDEM))
+	require.Nil(t, err)

 	// call IDriver methods
 	driverExt.TapXY(0.2, 0.5)
@@ -50,7 +55,7 @@ func TestDriverExt(t *testing.T) {
 	textRect, _ := driverExt.FindScreenText("hello")
 	t.Log(textRect)

-	err := driverExt.TapByCV(
+	err = driverExt.TapByCV(
 		option.WithScreenShotUITypes("deepseek_send"),
 		option.WithScope(0.8, 0.5, 1, 1))
 	assert.Nil(t, err)
@@ -249,3 +254,51 @@ func TestDriverExt_Action_Offset(t *testing.T) {
 		option.WithTapRandomRect(true))
 	assert.Nil(t, err)
 }
+
+func TestSaveImageWithCircle(t *testing.T) {
+	imgBytes, err := os.ReadFile("ai/testdata/llk_1.png")
+	require.NoError(t, err)
+	imgBuf := bytes.NewBuffer(imgBytes)
+
+	point := image.Point{X: 500, Y: 500}
+	outputPath := "ai/testdata/output.png"
+
+	err = SaveImageWithCircleMarker(imgBuf, point, outputPath)
+	require.NoError(t, err)
+
+	defer os.Remove(outputPath)
+}
+
+func TestSaveImageWithArrow(t *testing.T) {
+	imgBytes, err := os.ReadFile("ai/testdata/llk_1.png")
+	require.NoError(t, err)
+	imgBuf := bytes.NewBuffer(imgBytes)
+
+	from := image.Point{X: 500, Y: 500}
+	to := image.Point{X: 1000, Y: 1000}
+	outputPath := "ai/testdata/output.png"
+
+	err = SaveImageWithArrowMarker(imgBuf, from, to, outputPath)
+	require.NoError(t, err)
+
+	defer os.Remove(outputPath)
+}
+
+func TestMarkOperation(t *testing.T) {
+	driver := setupDriverExt(t)
+
+	opts := []option.ActionOption{option.WithMarkOperationEnabled(true)}
+
+	// tap point
+	err := driver.TapXY(0.5, 0.5, opts...)
+	assert.Nil(t, err)
+
+	err = driver.TapAbsXY(500, 800, opts...)
+	assert.Nil(t, err)
+
+	// swipe
+	err = driver.Swipe(0.2, 0.5, 0.8, 0.5, opts...)
+	assert.Nil(t, err)
+	err = driver.Swipe(0.3, 0.7, 0.3, 0.3, opts...)
+	assert.Nil(t, err)
+}
--- a/uixt/driver_handler.go
+++ b/uixt/driver_handler.go
@@ -0,0 +1,83 @@
+package uixt
+
+import (
+	"github.com/httprunner/httprunner/v5/uixt/option"
+	"github.com/rs/zerolog/log"
+)
+
+func handlerTapAbsXY(driver IDriver, rawX, rawY float64, opts ...option.ActionOption) (
+	x, y float64, err error) {
+
+	actionOptions := option.NewActionOptions(opts...)
+	x, y = actionOptions.ApplyTapOffset(rawX, rawY)
+
+	// mark UI operation
+	if actionOptions.MarkOperationEnabled {
+		if markErr := MarkUIOperation(driver, ACTION_TapAbsXY, []float64{x, y}); markErr != nil {
+			log.Warn().Err(markErr).Msg("Failed to mark tap operation")
+		}
+	}
+
+	return x, y, nil
+}
+
+func handlerDoubleTap(driver IDriver, rawX, rawY float64, opts ...option.ActionOption) (
+	x, y float64, err error) {
+
+	x, y, err = convertToAbsolutePoint(driver, rawX, rawY)
+	if err != nil {
+		return 0, 0, err
+	}
+
+	actionOptions := option.NewActionOptions(opts...)
+	x, y = actionOptions.ApplyTapOffset(x, y)
+
+	// mark UI operation
+	if actionOptions.MarkOperationEnabled {
+		if markErr := MarkUIOperation(driver, ACTION_DoubleTapXY, []float64{x, y}); markErr != nil {
+			log.Warn().Err(markErr).Msg("Failed to mark double tap operation")
+		}
+	}
+
+	return x, y, nil
+}
+
+func handlerDrag(driver IDriver, rawFomX, rawFromY, rawToX, rawToY float64, opts ...option.ActionOption) (
+	fromX, fromY, toX, toY float64, err error) {
+
+	actionOptions := option.NewActionOptions(opts...)
+	fromX, fromY, toX, toY, err = convertToAbsoluteCoordinates(driver, rawFomX, rawFromY, rawToX, rawToY)
+	if err != nil {
+		return 0, 0, 0, 0, err
+	}
+	fromX, fromY, toX, toY = actionOptions.ApplySwipeOffset(fromX, fromY, toX, toY)
+
+	// mark UI operation
+	if actionOptions.MarkOperationEnabled {
+		if markErr := MarkUIOperation(driver, ACTION_Drag, []float64{fromX, fromY, toX, toY}); markErr != nil {
+			log.Warn().Err(markErr).Msg("Failed to mark drag operation")
+		}
+	}
+
+	return fromX, fromY, toX, toY, nil
+}
+
+func handlerSwipe(driver IDriver, rawFomX, rawFromY, rawToX, rawToY float64, opts ...option.ActionOption) (
+	fromX, fromY, toX, toY float64, err error) {
+
+	actionOptions := option.NewActionOptions(opts...)
+	fromX, fromY, toX, toY, err = convertToAbsoluteCoordinates(driver, rawFomX, rawFromY, rawToX, rawToY)
+	if err != nil {
+		return 0, 0, 0, 0, err
+	}
+	fromX, fromY, toX, toY = actionOptions.ApplySwipeOffset(fromX, fromY, toX, toY)
+
+	// mark UI operation
+	if actionOptions.MarkOperationEnabled {
+		if markErr := MarkUIOperation(driver, ACTION_Swipe, []float64{fromX, fromY, toX, toY}); markErr != nil {
+			log.Warn().Err(markErr).Msg("Failed to mark swipe operation")
+		}
+	}
+
+	return fromX, fromY, toX, toY, nil
+}
--- a/uixt/driver_session.go
+++ b/uixt/driver_session.go
@@ -47,9 +47,9 @@ func NewDriverSession() *DriverSession {
 		client: &http.Client{
 			Timeout: timeout,
 		},
-		requests: make([]*DriverRequests, 0),
 		maxRetry: 5,
 	}
+	session.Reset()
 	return session
 }

@@ -66,10 +66,14 @@ type DriverSession struct {

 	// cache driver request and response
 	requests []*DriverRequests
+
+	// cache screenshot results
+	screenResults []*ScreenResult
 }

 func (s *DriverSession) Reset() {
 	s.requests = make([]*DriverRequests, 0)
+	s.screenResults = make([]*ScreenResult, 0)
 }

 func (s *DriverSession) SetBaseURL(baseUrl string) {
--- a/uixt/driver_utils.go
+++ b/uixt/driver_utils.go
@@ -117,11 +117,10 @@ func (dExt *XTDriver) GetData(withReset bool) map[string]interface{} {
 	session := dExt.GetSession()
 	data := map[string]interface{}{
 		"requests":       session.History(),
-		"screen_results": dExt.screenResults,
+		"screen_results": session.screenResults,
 	}
 	if withReset {
 		session.Reset()
-		dExt.screenResults = make([]*ScreenResult, 0)
 	}
 	return data
 }
--- a/uixt/harmony_driver_hdc.go
+++ b/uixt/harmony_driver_hdc.go
@@ -154,9 +154,14 @@ func (hd *HDCDriver) TapXY(x, y float64, opts ...option.ActionOption) error {

 func (hd *HDCDriver) TapAbsXY(x, y float64, opts ...option.ActionOption) error {
 	log.Info().Float64("x", x).Float64("y", y).Msg("HDCDriver.TapAbsXY")
-	actionOptions := option.NewActionOptions(opts...)
-	x, y = actionOptions.ApplyTapOffset(x, y)

+	var err error
+	x, y, err = handlerTapAbsXY(hd, x, y, opts...)
+	if err != nil {
+		return err
+	}
+
+	actionOptions := option.NewActionOptions(opts...)
 	if actionOptions.Identifier != "" {
 		startTime := int(time.Now().UnixMilli())
 		hd.points = append(hd.points, ExportPoint{Start: startTime, End: startTime + 100, Ext: actionOptions.Identifier, RunTime: 100})
@@ -182,13 +187,11 @@ func (hd *HDCDriver) Swipe(fromX, fromY, toX, toY float64, opts ...option.Action
 	log.Info().Float64("fromX", fromX).Float64("fromY", fromY).
 		Float64("toX", toX).Float64("toY", toY).Msg("HDCDriver.Swipe")
 	var err error
-	actionOptions := option.NewActionOptions(opts...)
-	fromX, fromY, toX, toY, err = convertToAbsoluteCoordinates(hd, fromX, fromY, toX, toY)
+	fromX, fromY, toX, toY, err = handlerSwipe(hd, fromX, fromY, toX, toY)
 	if err != nil {
 		return err
 	}
-	fromX, fromY, toX, toY = actionOptions.ApplySwipeOffset(fromX, fromY, toX, toY)
-
+	actionOptions := option.NewActionOptions(opts...)
 	duration := 200
 	if actionOptions.PressDuration > 0 {
 		duration = int(actionOptions.PressDuration * 1000)
--- a/uixt/harmony_test.go
+++ b/uixt/harmony_test.go
@@ -6,7 +6,7 @@ import (
 	"fmt"
 	"testing"

-	"github.com/httprunner/httprunner/v5/uixt/ai"
+	"github.com/httprunner/httprunner/v5/uixt/option"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 )
@@ -16,7 +16,9 @@ func setupHDCDriverExt(t *testing.T) *XTDriver {
 	require.Nil(t, err)
 	hdcDriver, err := NewHDCDriver(device)
 	require.Nil(t, err)
-	return NewXTDriver(hdcDriver, ai.WithCVService(ai.CVServiceTypeVEDEM))
+	driverExt, err := NewXTDriver(hdcDriver, option.WithCVService(option.CVServiceTypeVEDEM))
+	require.Nil(t, err)
+	return driverExt
 }

 func TestWindowSize(t *testing.T) {
--- a/uixt/ios_driver_wda.go
+++ b/uixt/ios_driver_wda.go
@@ -593,32 +593,40 @@ func (wd *WDADriver) TapXY(x, y float64, opts ...option.ActionOption) error {
 func (wd *WDADriver) TapAbsXY(x, y float64, opts ...option.ActionOption) error {
 	log.Info().Float64("x", x).Float64("y", y).Msg("WDADriver.TapAbsXY")
 	// [[FBRoute POST:@"/wda/tap/:uuid"] respondWithTarget:self action:@selector(handleTap:)]
-	actionOptions := option.NewActionOptions(opts...)
-	x, y = actionOptions.ApplyTapOffset(x, y)
+
+	x = wd.toScale(x)
+	y = wd.toScale(y)
+
+	var err error
+	x, y, err = handlerTapAbsXY(wd, x, y, opts...)
+	if err != nil {
+		return err
+	}
+
 	data := map[string]interface{}{
-		"x": wd.toScale(x),
-		"y": wd.toScale(y),
+		"x": x,
+		"y": y,
 	}
 	option.MergeOptions(data, opts...)

 	urlStr := fmt.Sprintf("/session/%s/wda/tap/0", wd.Session.ID)
-	_, err := wd.Session.POST(data, urlStr)
+	_, err = wd.Session.POST(data, urlStr)
 	return err
 }

 func (wd *WDADriver) DoubleTap(x, y float64, opts ...option.ActionOption) error {
 	log.Info().Float64("x", x).Float64("y", y).Msg("WDADriver.DoubleTap")
 	// [[FBRoute POST:@"/wda/doubleTap"] respondWithTarget:self action:@selector(handleDoubleTapCoordinate:)]
+
+	x = wd.toScale(x)
+	y = wd.toScale(y)
+
 	var err error
-	x, y, err = convertToAbsolutePoint(wd, x, y)
+	x, y, err = handlerDoubleTap(wd, x, y, opts...)
 	if err != nil {
 		return err
 	}

-	actionOptions := option.NewActionOptions(opts...)
-	x, y = actionOptions.ApplyTapOffset(x, y)
-	x = wd.toScale(x)
-	y = wd.toScale(y)
 	data := map[string]interface{}{
 		"x": x,
 		"y": y,
@@ -643,17 +651,17 @@ func (wd *WDADriver) Drag(fromX, fromY, toX, toY float64, opts ...option.ActionO
 	log.Info().Float64("fromX", fromX).Float64("fromY", fromY).
 		Float64("toX", toX).Float64("toY", toY).Msg("WDADriver.Drag")
 	// [[FBRoute POST:@"/wda/dragfromtoforduration"] respondWithTarget:self action:@selector(handleDragCoordinate:)]
-	var err error
-	fromX, fromY, toX, toY, err = convertToAbsoluteCoordinates(wd, fromX, fromY, toX, toY)
-	if err != nil {
-		return err
-	}
+
 	fromX = wd.toScale(fromX)
 	fromY = wd.toScale(fromY)
 	toX = wd.toScale(toX)
 	toY = wd.toScale(toY)
-	actionOptions := option.NewActionOptions(opts...)
-	fromX, fromY, toX, toY = actionOptions.ApplySwipeOffset(fromX, fromY, toX, toY)
+
+	var err error
+	fromX, fromY, toX, toY, err = handlerDrag(wd, fromX, fromY, toX, toY, opts...)
+	if err != nil {
+		return err
+	}

 	data := map[string]interface{}{
 		"fromX": math.Round(fromX*10) / 10,
@@ -970,7 +978,7 @@ func (wd *WDADriver) StartCaptureLog(identifier ...string) error {

 func (wd *WDADriver) PushImage(localPath string) error {
 	log.Info().Str("localPath", localPath).Msg("WDADriver.PushImage")
-	localFile, err := os.OpenFile(localPath, os.O_RDONLY, 0o600)
+	localFile, err := os.Open(localPath)
 	if err != nil {
 		return err
 	}
--- a/uixt/ios_test.go
+++ b/uixt/ios_test.go
@@ -10,7 +10,6 @@ import (
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"

-	"github.com/httprunner/httprunner/v5/uixt/ai"
 	"github.com/httprunner/httprunner/v5/uixt/option"
 	"github.com/httprunner/httprunner/v5/uixt/types"
 )
@@ -23,7 +22,9 @@ func setupWDADriverExt(t *testing.T) *XTDriver {
 	require.Nil(t, err)
 	driver, err := device.NewDriver()
 	require.Nil(t, err)
-	return NewXTDriver(driver, ai.WithCVService(ai.CVServiceTypeVEDEM))
+	driverExt, err := NewXTDriver(driver, option.WithCVService(option.CVServiceTypeVEDEM))
+	require.Nil(t, err)
+	return driverExt
 }

 func TestDevice_IOS_Install(t *testing.T) {
--- a/uixt/option/action.go
+++ b/uixt/option/action.go
@@ -132,6 +132,7 @@ func (o *ActionOptions) Options() []ActionOption {

 	options = append(options, o.GetScreenShotOptions()...)
 	options = append(options, o.GetScreenRecordOptions()...)
+	options = append(options, o.GetMarkOperationOptions()...)

 	return options
 }
--- a/uixt/option/ai.go
+++ b/uixt/option/ai.go
@@ -0,0 +1,43 @@
+package option
+
+func NewAIServiceOptions(opts ...AIServiceOption) *AIServiceOptions {
+	services := &AIServiceOptions{}
+	for _, option := range opts {
+		option(services)
+	}
+	return services
+}
+
+type AIServiceOptions struct {
+	CVService  CVServiceType
+	LLMService LLMServiceType
+}
+
+type AIServiceOption func(*AIServiceOptions)
+
+type CVServiceType string
+
+const (
+	CVServiceTypeVEDEM  CVServiceType = "vedem"
+	CVServiceTypeOpenCV CVServiceType = "opencv"
+)
+
+func WithCVService(service CVServiceType) AIServiceOption {
+	return func(opts *AIServiceOptions) {
+		opts.CVService = service
+	}
+}
+
+type LLMServiceType string
+
+const (
+	LLMServiceTypeUITARS LLMServiceType = "ui-tars"
+	LLMServiceTypeGPT    LLMServiceType = "gpt"
+	LLMServiceTypeQwenVL LLMServiceType = "qwen-vl"
+)
+
+func WithLLMService(modelType LLMServiceType) AIServiceOption {
+	return func(opts *AIServiceOptions) {
+		opts.LLMService = modelType
+	}
+}
--- a/uixt/option/screen.go
+++ b/uixt/option/screen.go
@@ -8,6 +8,7 @@ type ScreenOptions struct {
 	ScreenShotOptions
 	ScreenRecordOptions
 	ScreenFilterOptions
+	MarkOperationOptions
 }

 type ScreenShotOptions struct {
@@ -273,3 +274,29 @@ func WithIndex(index int) ActionOption {
 		o.Index = index
 	}
 }
+
+// MarkOperationOptions contains options for marking UI operations
+type MarkOperationOptions struct {
+	// mark UI operation, enable/disable UI operation marking
+	MarkOperationEnabled bool `json:"mark_operation_enabled,omitempty" yaml:"mark_operation_enabled,omitempty"`
+}
+
+func (o *MarkOperationOptions) GetMarkOperationOptions() []ActionOption {
+	options := make([]ActionOption, 0)
+	if o == nil {
+		return options
+	}
+
+	if o.MarkOperationEnabled {
+		options = append(options, WithMarkOperationEnabled(true))
+	}
+
+	return options
+}
+
+// WithMarkOperationEnabled enables or disables UI operation marking
+func WithMarkOperationEnabled(enabled bool) ActionOption {
+	return func(o *ActionOptions) {
+		o.MarkOperationEnabled = enabled
+	}
+}