diff --git a/src/renderer/src/components/MinApp/index.tsx b/src/renderer/src/components/MinApp/index.tsx
index 109f28b9..f781eb1f 100644
--- a/src/renderer/src/components/MinApp/index.tsx
+++ b/src/renderer/src/components/MinApp/index.tsx
@@ -54,6 +54,7 @@ const PopupContainer: React.FC<Props> = ({ app, resolve }) => {
     const newPinned = isPinned ? pinned.filter((item) => item.id !== app.id) : [...pinned, app]
     updatePinnedMinapps(newPinned)
   }
+
   const Title = () => {
     return (
       <TitleContainer style={{ justifyContent: 'space-between' }}>
@@ -63,7 +64,7 @@ const PopupContainer: React.FC<Props> = ({ app, resolve }) => {
             <ReloadOutlined />
           </Button>
           <Button onClick={onTogglePin} className={isPinned ? 'pinned' : ''}>
-            <PushpinOutlined />
+            <PushpinOutlined style={{ fontSize: 16 }} />
           </Button>
           {canOpenExternalLink && (
             <Button onClick={onOpenLink}>
diff --git a/src/renderer/src/config/minapps.ts b/src/renderer/src/config/minapps.ts
index f1aceeed..2b99fb20 100644
--- a/src/renderer/src/config/minapps.ts
+++ b/src/renderer/src/config/minapps.ts
@@ -1,38 +1,38 @@
-import BaiduAiAppLogo from '@renderer/assets/images/apps/baidu-ai.png'
-import BaicuanAppLogo from '@renderer/assets/images/apps/baixiaoying.webp'
-import BoltAppLogo from '@renderer/assets/images/apps/bolt.svg'
-import DevvAppLogo from '@renderer/assets/images/apps/devv.png'
-import DoubaoAppLogo from '@renderer/assets/images/apps/doubao.png'
-import DuckDuckGoAppLogo from '@renderer/assets/images/apps/duckduckgo.webp'
-import FeloAppLogo from '@renderer/assets/images/apps/felo.png'
-import GeminiAppLogo from '@renderer/assets/images/apps/gemini.png'
-import GensparkLogo from '@renderer/assets/images/apps/genspark.jpg'
-import GithubCopilotLogo from '@renderer/assets/images/apps/github-copilot.webp'
-import GrokAppLogo from '@renderer/assets/images/apps/grok.png'
-import HikaLogo from '@renderer/assets/images/apps/hika.webp'
-import HuggingChatLogo from '@renderer/assets/images/apps/huggingchat.svg'
-import KimiAppLogo from '@renderer/assets/images/apps/kimi.jpg'
-import MetasoAppLogo from '@renderer/assets/images/apps/metaso.webp'
-import NamiAiSearchLogo from '@renderer/assets/images/apps/nm.webp'
-import PerplexityAppLogo from '@renderer/assets/images/apps/perplexity.webp'
-import PoeAppLogo from '@renderer/assets/images/apps/poe.webp'
-import ZhipuProviderLogo from '@renderer/assets/images/apps/qingyan.png'
-import QwenlmAppLogo from '@renderer/assets/images/apps/qwenlm.webp'
-import SensetimeAppLogo from '@renderer/assets/images/apps/sensetime.png'
-import SparkDeskAppLogo from '@renderer/assets/images/apps/sparkdesk.png'
-import ThinkAnyLogo from '@renderer/assets/images/apps/thinkany.webp'
-import TiangongAiLogo from '@renderer/assets/images/apps/tiangong.png'
-import WanZhiAppLogo from '@renderer/assets/images/apps/wanzhi.jpg'
-import TencentYuanbaoAppLogo from '@renderer/assets/images/apps/yuanbao.png'
-import YuewenAppLogo from '@renderer/assets/images/apps/yuewen.png'
-import ZhihuAppLogo from '@renderer/assets/images/apps/zhihu.png'
-import ClaudeAppLogo from '@renderer/assets/images/models/claude.png'
-import HailuoModelLogo from '@renderer/assets/images/models/hailuo.png'
-import QwenModelLogo from '@renderer/assets/images/models/qwen.png'
-import DeepSeekProviderLogo from '@renderer/assets/images/providers/deepseek.png'
-import GroqProviderLogo from '@renderer/assets/images/providers/groq.png'
-import OpenAiProviderLogo from '@renderer/assets/images/providers/openai.png'
-import SiliconFlowProviderLogo from '@renderer/assets/images/providers/silicon.png'
+import BaiduAiAppLogo from '@renderer/assets/images/apps/baidu-ai.png?url'
+import BaicuanAppLogo from '@renderer/assets/images/apps/baixiaoying.webp?url'
+import BoltAppLogo from '@renderer/assets/images/apps/bolt.svg?url'
+import DevvAppLogo from '@renderer/assets/images/apps/devv.png?url'
+import DoubaoAppLogo from '@renderer/assets/images/apps/doubao.png?url'
+import DuckDuckGoAppLogo from '@renderer/assets/images/apps/duckduckgo.webp?url'
+import FeloAppLogo from '@renderer/assets/images/apps/felo.png?url'
+import GeminiAppLogo from '@renderer/assets/images/apps/gemini.png?url'
+import GensparkLogo from '@renderer/assets/images/apps/genspark.jpg?url'
+import GithubCopilotLogo from '@renderer/assets/images/apps/github-copilot.webp?url'
+import GrokAppLogo from '@renderer/assets/images/apps/grok.png?url'
+import HikaLogo from '@renderer/assets/images/apps/hika.webp?url'
+import HuggingChatLogo from '@renderer/assets/images/apps/huggingchat.svg?url'
+import KimiAppLogo from '@renderer/assets/images/apps/kimi.jpg?url'
+import MetasoAppLogo from '@renderer/assets/images/apps/metaso.webp?url'
+import NamiAiSearchLogo from '@renderer/assets/images/apps/nm.webp?url'
+import PerplexityAppLogo from '@renderer/assets/images/apps/perplexity.webp?url'
+import PoeAppLogo from '@renderer/assets/images/apps/poe.webp?url'
+import ZhipuProviderLogo from '@renderer/assets/images/apps/qingyan.png?url'
+import QwenlmAppLogo from '@renderer/assets/images/apps/qwenlm.webp?url'
+import SensetimeAppLogo from '@renderer/assets/images/apps/sensetime.png?url'
+import SparkDeskAppLogo from '@renderer/assets/images/apps/sparkdesk.png?url'
+import ThinkAnyLogo from '@renderer/assets/images/apps/thinkany.webp?url'
+import TiangongAiLogo from '@renderer/assets/images/apps/tiangong.png?url'
+import WanZhiAppLogo from '@renderer/assets/images/apps/wanzhi.jpg?url'
+import TencentYuanbaoAppLogo from '@renderer/assets/images/apps/yuanbao.png?url'
+import YuewenAppLogo from '@renderer/assets/images/apps/yuewen.png?url'
+import ZhihuAppLogo from '@renderer/assets/images/apps/zhihu.png?url'
+import ClaudeAppLogo from '@renderer/assets/images/models/claude.png?url'
+import HailuoModelLogo from '@renderer/assets/images/models/hailuo.png?url'
+import QwenModelLogo from '@renderer/assets/images/models/qwen.png?url'
+import DeepSeekProviderLogo from '@renderer/assets/images/providers/deepseek.png?url'
+import GroqProviderLogo from '@renderer/assets/images/providers/groq.png?url'
+import OpenAiProviderLogo from '@renderer/assets/images/providers/openai.png?url'
+import SiliconFlowProviderLogo from '@renderer/assets/images/providers/silicon.png?url'
 import MinApp from '@renderer/components/MinApp'
 import { MinAppType } from '@renderer/types'
 
diff --git a/src/renderer/src/pages/apps/App.tsx b/src/renderer/src/pages/apps/App.tsx
index a195d380..e5592779 100644
--- a/src/renderer/src/pages/apps/App.tsx
+++ b/src/renderer/src/pages/apps/App.tsx
@@ -30,6 +30,7 @@ const App: FC<Props> = ({ app, onClick, size = 60 }) => {
       key: 'togglePin',
       label: isPinned ? t('minapp.sidebar.remove.title') : t('minapp.sidebar.add.title'),
       onClick: () => {
+        console.debug('togglePin', app)
         const newPinned = isPinned ? pinned.filter((item) => item.id !== app.id) : [...(pinned || []), app]
         updatePinnedMinapps(newPinned)
       }
diff --git a/src/renderer/src/pages/apps/AppsPage.tsx b/src/renderer/src/pages/apps/AppsPage.tsx
index 107cc875..912ca371 100644
--- a/src/renderer/src/pages/apps/AppsPage.tsx
+++ b/src/renderer/src/pages/apps/AppsPage.tsx
@@ -15,8 +15,6 @@ const AppsPage: FC = () => {
   const [search, setSearch] = useState('')
   const { minapps } = useMinapps()
 
-  console.debug('minapps', minapps)
-
   const filteredApps = search
     ? minapps.filter(
         (app) => app.name.toLowerCase().includes(search.toLowerCase()) || app.url.includes(search.toLowerCase())
diff --git a/src/renderer/src/providers/OpenAIProvider.ts b/src/renderer/src/providers/OpenAIProvider.ts
index 4730f53d..87cbfcf8 100644
--- a/src/renderer/src/providers/OpenAIProvider.ts
+++ b/src/renderer/src/providers/OpenAIProvider.ts
@@ -46,33 +46,6 @@ export default class OpenAIProvider extends BaseProvider {
     return providers.includes(this.provider.id)
   }
 
-  private async uploadImageToQwenLM(image_file: Buffer, file_name: string, mime: string): Promise<string> {
-    try {
-      // 创建 FormData
-      const formData = new FormData()
-      formData.append('file', new Blob([image_file], { type: mime }), file_name)
-
-      // 发送上传请求
-      const response = await fetch(`${this.provider.apiHost}v1/files/`, {
-        method: 'POST',
-        headers: {
-          Authorization: `Bearer ${this.apiKey}`
-        },
-        body: formData
-      })
-
-      if (!response.ok) {
-        throw new Error('Failed to upload image to QwenLM')
-      }
-
-      const data = await response.json()
-      return data.id
-    } catch (error) {
-      console.error('Error uploading image to QwenLM:', error)
-      throw error
-    }
-  }
-
   private async getMessageParam(
     message: Message,
     model: Model
@@ -121,34 +94,6 @@ export default class OpenAIProvider extends BaseProvider {
       }
     ]
 
-    //QwenLM上传图片
-    if (this.provider.id === 'qwenlm') {
-      const qwenlm_image_url: { type: string; image: string }[] = []
-
-      for (const file of message.files || []) {
-        if (file.type === FileTypes.IMAGE && isVision) {
-          const image = await window.api.file.binaryFile(file.id + file.ext)
-
-          const imageId = await this.uploadImageToQwenLM(image.data, file.origin_name, image.mime)
-          qwenlm_image_url.push({
-            type: 'image',
-            image: imageId
-          })
-        }
-        if ([FileTypes.TEXT, FileTypes.DOCUMENT].includes(file.type)) {
-          const fileContent = await (await window.api.file.read(file.id + file.ext)).trim()
-          parts.push({
-            type: 'text',
-            text: file.origin_name + '\n' + fileContent
-          })
-        }
-      }
-      return {
-        role: message.role,
-        content: [...parts, ...qwenlm_image_url]
-      } as ChatCompletionMessageParam
-    }
-
     for (const file of message.files || []) {
       if (file.type === FileTypes.IMAGE && isVision) {
         const image = await window.api.file.base64Image(file.id + file.ext)
@@ -183,10 +128,6 @@ export default class OpenAIProvider extends BaseProvider {
     const _messages = filterContextMessages(takeRight(messages, contextCount + 1))
     onFilterMessages(_messages)
 
-    if (this.provider.id === 'qwenlm' && _messages[0]?.role !== 'user') {
-      userMessages.push({ role: 'user', content: '' })
-    }
-
     for (const message of _messages) {
       userMessages.push(await this.getMessageParam(message, model))
     }
@@ -231,40 +172,6 @@ export default class OpenAIProvider extends BaseProvider {
       })
     }
 
-    // 处理QwenLM的流式输出
-    if (this.provider.id === 'qwenlm') {
-      let accumulatedText = ''
-      for await (const chunk of stream) {
-        if (window.keyv.get(EVENT_NAMES.CHAT_COMPLETION_PAUSED)) {
-          break
-        }
-        if (time_first_token_millsec == 0) {
-          time_first_token_millsec = new Date().getTime() - start_time_millsec
-        }
-
-        // 获取当前块的完整内容
-        const currentContent = chunk.choices[0]?.delta?.content || ''
-
-        // 如果内容与累积的内容不同，则只发送增量部分
-        if (currentContent !== accumulatedText) {
-          const deltaText = currentContent.slice(accumulatedText.length)
-          accumulatedText = currentContent // 更新累积的文本
-
-          const time_completion_millsec = new Date().getTime() - start_time_millsec
-          onChunk({
-            text: deltaText,
-            usage: chunk.usage,
-            metrics: {
-              completion_tokens: chunk.usage?.completion_tokens,
-              time_completion_millsec,
-              time_first_token_millsec
-            }
-          })
-        }
-      }
-      return
-    }
-
     for await (const chunk of stream) {
       if (window.keyv.get(EVENT_NAMES.CHAT_COMPLETION_PAUSED)) {
         break
diff --git a/src/renderer/src/providers/ProviderFactory.ts b/src/renderer/src/providers/ProviderFactory.ts
index c730020a..ade11c0d 100644
--- a/src/renderer/src/providers/ProviderFactory.ts
+++ b/src/renderer/src/providers/ProviderFactory.ts
@@ -4,6 +4,7 @@ import AnthropicProvider from './AnthropicProvider'
 import BaseProvider from './BaseProvider'
 import GeminiProvider from './GeminiProvider'
 import OpenAIProvider from './OpenAIProvider'
+import QwenLMProvider from './QwenLMProvider'
 
 export default class ProviderFactory {
   static create(provider: Provider): BaseProvider {
@@ -12,6 +13,8 @@ export default class ProviderFactory {
         return new AnthropicProvider(provider)
       case 'gemini':
         return new GeminiProvider(provider)
+      case 'qwenlm':
+        return new QwenLMProvider(provider)
       default:
         return new OpenAIProvider(provider)
     }
diff --git a/src/renderer/src/providers/QwenLMProvider.ts b/src/renderer/src/providers/QwenLMProvider.ts
new file mode 100644
index 00000000..f701bde1
--- /dev/null
+++ b/src/renderer/src/providers/QwenLMProvider.ts
@@ -0,0 +1,160 @@
+import { getOpenAIWebSearchParams, isVisionModel } from '@renderer/config/models'
+import { getAssistantSettings, getDefaultModel } from '@renderer/services/AssistantService'
+import { EVENT_NAMES } from '@renderer/services/EventService'
+import { filterContextMessages } from '@renderer/services/MessagesService'
+import { FileTypes, Message, Model, Provider } from '@renderer/types'
+import { takeRight } from 'lodash'
+import OpenAI from 'openai'
+import { ChatCompletionContentPart, ChatCompletionMessageParam } from 'openai/resources'
+
+import { CompletionsParams } from '.'
+import OpenAIProvider from './OpenAIProvider'
+
+class QwenLMProvider extends OpenAIProvider {
+  constructor(provider: Provider) {
+    super(provider)
+  }
+
+  private async getMessageParams(
+    message: Message,
+    model: Model
+  ): Promise<OpenAI.Chat.Completions.ChatCompletionMessageParam> {
+    const isVision = isVisionModel(model)
+    const content = await this.getMessageContent(message)
+
+    if (!message.files) {
+      return {
+        role: message.role,
+        content
+      }
+    }
+
+    const parts: ChatCompletionContentPart[] = [
+      {
+        type: 'text',
+        text: content
+      }
+    ]
+
+    const qwenlm_image_url: { type: string; image: string }[] = []
+
+    for (const file of message.files || []) {
+      if (file.type === FileTypes.IMAGE && isVision) {
+        const image = await window.api.file.binaryFile(file.id + file.ext)
+
+        const imageId = await this.uploadImageToQwenLM(image.data, file.origin_name, image.mime)
+        qwenlm_image_url.push({
+          type: 'image',
+          image: imageId
+        })
+      }
+      if ([FileTypes.TEXT, FileTypes.DOCUMENT].includes(file.type)) {
+        const fileContent = await (await window.api.file.read(file.id + file.ext)).trim()
+        parts.push({
+          type: 'text',
+          text: file.origin_name + '\n' + fileContent
+        })
+      }
+    }
+
+    return {
+      role: message.role,
+      content: [...parts, ...qwenlm_image_url]
+    } as ChatCompletionMessageParam
+  }
+
+  private async uploadImageToQwenLM(image_file: Buffer, file_name: string, mime: string): Promise<string> {
+    try {
+      // 创建 FormData
+      const formData = new FormData()
+      formData.append('file', new Blob([image_file], { type: mime }), file_name)
+
+      // 发送上传请求
+      const response = await fetch(`${this.provider.apiHost}v1/files/`, {
+        method: 'POST',
+        headers: {
+          Authorization: `Bearer ${this.apiKey}`
+        },
+        body: formData
+      })
+
+      if (!response.ok) {
+        throw new Error('Failed to upload image to QwenLM')
+      }
+
+      const data = await response.json()
+      return data.id
+    } catch (error) {
+      console.error('Error uploading image to QwenLM:', error)
+      throw error
+    }
+  }
+
+  async completions({ messages, assistant, onChunk, onFilterMessages }: CompletionsParams): Promise<void> {
+    const defaultModel = getDefaultModel()
+    const model = assistant.model || defaultModel
+    const { contextCount, maxTokens } = getAssistantSettings(assistant)
+
+    const systemMessage = assistant.prompt ? { role: 'system', content: assistant.prompt } : undefined
+    const userMessages: ChatCompletionMessageParam[] = []
+
+    const _messages = filterContextMessages(takeRight(messages, contextCount + 1))
+    onFilterMessages(_messages)
+
+    if (_messages[0]?.role !== 'user') {
+      userMessages.push({ role: 'user', content: '' })
+    }
+
+    for (const message of _messages) {
+      userMessages.push(await this.getMessageParams(message, model))
+    }
+
+    let time_first_token_millsec = 0
+    const start_time_millsec = new Date().getTime()
+
+    // @ts-ignore key is not typed
+    const stream = await this.sdk.chat.completions.create({
+      model: model.id,
+      messages: [systemMessage, ...userMessages].filter(Boolean) as ChatCompletionMessageParam[],
+      temperature: assistant?.settings?.temperature,
+      top_p: assistant?.settings?.topP,
+      max_tokens: maxTokens,
+      stream: true,
+      ...(assistant.enableWebSearch ? getOpenAIWebSearchParams(model) : {}),
+      ...this.getCustomParameters(assistant)
+    })
+
+    let accumulatedText = ''
+
+    for await (const chunk of stream) {
+      if (window.keyv.get(EVENT_NAMES.CHAT_COMPLETION_PAUSED)) {
+        break
+      }
+      if (time_first_token_millsec == 0) {
+        time_first_token_millsec = new Date().getTime() - start_time_millsec
+      }
+
+      // 获取当前块的完整内容
+      const currentContent = chunk.choices[0]?.delta?.content || ''
+
+      // 如果内容与累积的内容不同，则只发送增量部分
+      if (currentContent !== accumulatedText) {
+        const deltaText = currentContent.slice(accumulatedText.length)
+        accumulatedText = currentContent // 更新累积的文本
+
+        const time_completion_millsec = new Date().getTime() - start_time_millsec
+        onChunk({
+          text: deltaText,
+          usage: chunk.usage,
+          metrics: {
+            completion_tokens: chunk.usage?.completion_tokens,
+            time_completion_millsec,
+            time_first_token_millsec
+          }
+        })
+      }
+    }
+  }
+}
+
+export default QwenLMProvider
diff --git a/src/renderer/src/types/index.ts b/src/renderer/src/types/index.ts
index 78d7da88..4b5618f1 100644
--- a/src/renderer/src/types/index.ts
+++ b/src/renderer/src/types/index.ts
@@ -101,7 +101,7 @@ export type Provider = {
   isSystem?: boolean
 }
 
-export type ProviderType = 'openai' | 'anthropic' | 'gemini'
+export type ProviderType = 'openai' | 'anthropic' | 'gemini' | 'qwenlm'
 
 export type ModelType = 'text' | 'vision' | 'embedding'