refactor: migrate to v5 patch-1

2025-12-28 05:11:24 +08:00 · 2025-07-06 04:25:11 +08:00 · 2025-07-06 04:25:11 +08:00 · 3e5969b97c
commit 3e5969b97c
parent cd42410d70
16 changed files with 223 additions and 245 deletions
--- a/packages/aiCore/package.json
+++ b/packages/aiCore/package.json
@ -46,7 +46,8 @@
    "@openrouter/ai-sdk-provider": "^0.7.2",
    "ai": "5.0.0-beta.7",
    "anthropic-vertex-ai": "^1.0.2",
-    "ollama-ai-provider": "^1.2.0"
+    "ollama-ai-provider": "^1.2.0",
+    "zod": "^3.25.0"
  },
  "peerDependenciesMeta": {
    "@ai-sdk/amazon-bedrock": {
--- a/packages/aiCore/src/core/index.ts
+++ b/packages/aiCore/src/core/index.ts
@ -19,5 +19,6 @@ export {
 } from './models'

 // 执行管理
+export type { MCPRequestContext } from './plugins/built-in/mcpPromptPlugin'
 export type { ExecutionOptions, ExecutorConfig } from './runtime'
 export { createExecutor, createOpenAICompatibleExecutor } from './runtime'
--- a/packages/aiCore/src/core/options/types.ts
+++ b/packages/aiCore/src/core/options/types.ts
@ -1,11 +1,11 @@
 import { type AnthropicProviderOptions } from '@ai-sdk/anthropic'
 import { type GoogleGenerativeAIProviderOptions } from '@ai-sdk/google'
 import { type OpenAIResponsesProviderOptions } from '@ai-sdk/openai'
-import { type LanguageModelV1ProviderMetadata } from '@ai-sdk/provider'
+import { type SharedV2ProviderMetadata } from '@ai-sdk/provider'

 import { type OpenRouterProviderOptions } from './openrouter'

-export type ProviderOptions<T extends keyof LanguageModelV1ProviderMetadata> = LanguageModelV1ProviderMetadata[T]
+export type ProviderOptions<T extends keyof SharedV2ProviderMetadata> = SharedV2ProviderMetadata[T]

 /**
 * 供应商选项类型，如果map中没有，说明没有约束
@ -28,4 +28,4 @@ export type TypedProviderOptions = {
  [K in keyof ProviderOptionsMap]?: ProviderOptionsMap[K]
 } & {
  [K in string]?: Record<string, any>
-} & LanguageModelV1ProviderMetadata
+} & SharedV2ProviderMetadata
--- a/packages/aiCore/src/core/plugins/built-in/mcpPromptPlugin.ts
+++ b/packages/aiCore/src/core/plugins/built-in/mcpPromptPlugin.ts
@ -3,8 +3,7 @@
 * 为不支持原生 Function Call 的模型提供 prompt 方式的工具调用
 * 内置默认逻辑，支持自定义覆盖
 */
-import type { ToolSet } from 'ai'
-import { ToolExecutionError } from 'ai'
+import type { ModelMessage, TextStreamPart, ToolErrorUnion, ToolSet } from 'ai'

 import { definePlugin } from '../index'
 import type { AiRequestContext } from '../types'
@ -44,17 +43,17 @@ export interface MCPPromptConfig {
  // 是否启用（用于运行时开关）
  enabled?: boolean
  // 自定义系统提示符构建函数（可选，有默认实现）
-  buildSystemPrompt?: (userSystemPrompt: string, tools: ToolSet) => Promise<string>
+  buildSystemPrompt?: (userSystemPrompt: string, tools: ToolSet) => string
  // 自定义工具解析函数（可选，有默认实现）
  parseToolUse?: (content: string, tools: ToolSet) => ToolUseResult[]
-  createSystemMessage?: (systemPrompt: string, originalParams: any, context: MCPRequestContext) => string | null
+  createSystemMessage?: (systemPrompt: string, originalParams: any, context: AiRequestContext) => string | null
 }

 /**
 * 扩展的 AI 请求上下文，支持 MCP 工具存储
 */
-interface MCPRequestContext extends AiRequestContext {
-  mcpTools?: ToolSet
+export interface MCPRequestContext extends AiRequestContext {
+  mcpTools: ToolSet
 }

 /**
@ -201,7 +200,7 @@ function buildAvailableTools(tools: ToolSet): string {
  <name>${toolName}</name>
  <description>${tool.description || ''}</description>
  <arguments>
-    ${tool.parameters ? JSON.stringify(tool.parameters) : ''}
+    ${tool.inputSchema ? JSON.stringify(tool.inputSchema) : ''}
  </arguments>
 </tool>
 `
@ -215,7 +214,7 @@ ${availableTools}
 /**
 * 默认的系统提示符构建函数（提取自 Cherry Studio）
 */
-async function defaultBuildSystemPrompt(userSystemPrompt: string, tools: ToolSet): Promise<string> {
+function defaultBuildSystemPrompt(userSystemPrompt: string, tools: ToolSet): string {
  const availableTools = buildAvailableTools(tools)

  const fullPrompt = DEFAULT_SYSTEM_PROMPT.replace('{{ TOOL_USE_EXAMPLES }}', DEFAULT_TOOL_USE_EXAMPLES)
@ -291,8 +290,7 @@ export const createMCPPromptPlugin = (config: MCPPromptConfig = {}) => {

  return definePlugin({
    name: 'built-in:mcp-prompt',
-
-    transformParams: async (params: any, context: MCPRequestContext) => {
+    transformParams: (params: any, context: AiRequestContext) => {
      if (!enabled || !params.tools || typeof params.tools !== 'object') {
        return params
      }
@ -303,7 +301,7 @@ export const createMCPPromptPlugin = (config: MCPPromptConfig = {}) => {

      // 构建系统提示符
      const userSystemPrompt = typeof params.system === 'string' ? params.system : ''
-      const systemPrompt = await buildSystemPrompt(userSystemPrompt, params.tools)
+      const systemPrompt = buildSystemPrompt(userSystemPrompt, params.tools)
      let systemMessage: string | null = systemPrompt
      console.log('config.context', context)
      if (config.createSystemMessage) {
@ -320,25 +318,30 @@ export const createMCPPromptPlugin = (config: MCPPromptConfig = {}) => {
      console.log('transformedParams', transformedParams)
      return transformedParams
    },
-
-    // 流式处理：监听 step-finish 事件并处理工具调用
-    transformStream: (_, context: MCPRequestContext) => () => {
+    transformStream: (_: any, context: AiRequestContext) => () => {
      let textBuffer = ''
+      let stepId = ''
      let executedResults: { toolCallId: string; toolName: string; result: any; isError?: boolean }[] = []
-
-      return new TransformStream<any>({
-        async transform(chunk, controller) {
+      if (!context.mcpTools) {
+        throw new Error('No tools available')
+      }
+      type TOOLS = NonNullable<typeof context.mcpTools>
+      return new TransformStream<TextStreamPart<TOOLS>, TextStreamPart<TOOLS>>({
+        async transform(
+          chunk: TextStreamPart<TOOLS>,
+          controller: TransformStreamDefaultController<TextStreamPart<TOOLS>>
+        ) {
          // console.log('chunk', chunk)
          // 收集文本内容
-          if (chunk.type === 'text-delta') {
-            textBuffer += chunk.textDelta || ''
+          if (chunk.type === 'text') {
+            textBuffer += chunk.text || ''
+            stepId = chunk.id || ''
            // console.log('textBuffer', textBuffer)
            controller.enqueue(chunk)
            return
          }

-          // 监听 step-finish 事件
-          if (chunk.type === 'step-finish' || chunk.type === 'finish') {
+          if (chunk.type === 'finish-step') {
            // console.log('[MCP Prompt Stream] Received step-finish, checking for tool use...')

            // 从 context 获取工具信息
@ -364,17 +367,11 @@ export const createMCPPromptPlugin = (config: MCPPromptConfig = {}) => {

            // console.log('[MCP Prompt Stream] Found valid tool uses:', validToolUses.length)

-            // 修改 step-finish 事件，标记为工具调用
-            if (chunk.type !== 'finish') {
-              controller.enqueue({
-                ...chunk,
-                finishReason: 'tool-call'
-              })
-            }
-
            // 发送 step-start 事件（工具调用步骤开始）
            controller.enqueue({
-              type: 'step-start'
+              type: 'start-step',
+              request: {},
+              warnings: []
            })

            // 执行工具调用
@ -392,7 +389,7 @@ export const createMCPPromptPlugin = (config: MCPPromptConfig = {}) => {
                  type: 'tool-call',
                  toolCallId: toolUse.id,
                  toolName: toolUse.toolName,
-                  args: toolUse.arguments
+                  input: tool.inputSchema
                })

                const result = await tool.execute(toolUse.arguments, {
@ -406,8 +403,8 @@ export const createMCPPromptPlugin = (config: MCPPromptConfig = {}) => {
                  type: 'tool-result',
                  toolCallId: toolUse.id,
                  toolName: toolUse.toolName,
-                  args: toolUse.arguments,
-                  result
+                  input: toolUse.arguments,
+                  output: result
                })

                executedResults.push({
@ -420,39 +417,36 @@ export const createMCPPromptPlugin = (config: MCPPromptConfig = {}) => {
                console.error(`[MCP Prompt Stream] Tool execution failed: ${toolUse.toolName}`, error)

                // 使用 AI SDK 标准错误格式
-                const toolError = new ToolExecutionError({
-                  toolName: toolUse.toolName,
-                  toolArgs: toolUse.arguments,
+                const toolError: ToolErrorUnion<typeof context.mcpTools> = {
+                  type: 'tool-error',
                  toolCallId: toolUse.id,
-                  message: `Tool execution failed: ${error instanceof Error ? error.message : String(error)}`,
-                  cause: error instanceof Error ? error : undefined
-                })
+                  toolName: toolUse.toolName,
+                  input: toolUse.arguments,
+                  error: error instanceof Error ? error.message : String(error)
+                }
+
+                controller.enqueue(toolError)

                // 发送标准错误事件
                controller.enqueue({
                  type: 'error',
-                  error: {
-                    message: toolError.message,
-                    name: toolError.name,
-                    toolName: toolError.toolName,
-                    toolCallId: toolError.toolCallId
-                  }
+                  error: toolError.error
                })

-                // 发送 tool-result 错误事件
-                controller.enqueue({
-                  type: 'tool-result',
-                  toolCallId: toolUse.id,
-                  toolName: toolUse.toolName,
-                  args: toolUse.arguments,
-                  isError: true,
-                  result: toolError.message
-                })
+                // // 发送 tool-result 错误事件
+                // controller.enqueue({
+                //   type: 'tool-result',
+                //   toolCallId: toolUse.id,
+                //   toolName: toolUse.toolName,
+                //   args: toolUse.arguments,
+                //   isError: true,
+                //   result: toolError.message
+                // })

                executedResults.push({
                  toolCallId: toolUse.id,
                  toolName: toolUse.toolName,
-                  result: toolError.message,
+                  result: toolError.error,
                  isError: true
                })
              }
@ -460,8 +454,11 @@ export const createMCPPromptPlugin = (config: MCPPromptConfig = {}) => {

            // 发送最终的 step-finish 事件
            controller.enqueue({
-              type: 'step-finish',
-              finishReason: 'tool-call'
+              type: 'finish-step',
+              finishReason: 'tool-calls',
+              response: chunk.response,
+              usage: chunk.usage,
+              providerMetadata: chunk.providerMetadata
            })

            // 递归调用逻辑
@ -481,7 +478,7 @@ export const createMCPPromptPlugin = (config: MCPPromptConfig = {}) => {
                .join('\n\n')
              // console.log('context.originalParams.messages', context.originalParams.messages)
              // 构建新的对话消息
-              const newMessages = [
+              const newMessages: ModelMessage[] = [
                ...(context.originalParams.messages || []),
                {
                  role: 'assistant',
@ -540,8 +537,9 @@ export const createMCPPromptPlugin = (config: MCPPromptConfig = {}) => {

                // 继续发送文本增量，保持流的连续性
                controller.enqueue({
-                  type: 'text-delta',
-                  textDelta: '\n\n[工具执行后递归调用失败，继续对话...]'
+                  type: 'text',
+                  id: stepId,
+                  text: '\n\n[工具执行后递归调用失败，继续对话...]'
                })
              }
            }
--- a/packages/aiCore/src/core/plugins/index.ts
+++ b/packages/aiCore/src/core/plugins/index.ts
@ -1,12 +1,13 @@
 // 核心类型和接口
 export type { AiPlugin, AiRequestContext, HookResult, HookType, PluginManagerConfig } from './types'
+import { ProviderId } from '../providers/registry'
 import type { AiPlugin, AiRequestContext } from './types'

 // 插件管理器
 export { PluginManager } from './manager'

 // 工具函数
-export function createContext(providerId: string, modelId: string, originalParams: any): AiRequestContext {
+export function createContext(providerId: ProviderId, modelId: string, originalParams: any): AiRequestContext {
  return {
    providerId,
    modelId,
--- a/packages/aiCore/src/core/plugins/types.ts
+++ b/packages/aiCore/src/core/plugins/types.ts
@ -20,6 +20,7 @@ export interface AiRequestContext {
  requestId: string
  recursiveCall: RecursiveCallFn
  isRecursiveCall?: boolean
+  mcpTools?: ToolSet
  [key: string]: any
 }

@ -47,7 +48,7 @@ export interface AiPlugin {
  transformStream?: (
    params: any,
    context: AiRequestContext
-  ) => <TOOLS extends ToolSet>(options: {
+  ) => <TOOLS extends ToolSet>(options?: {
    tools: TOOLS
    stopStream: () => void
  }) => TransformStream<TextStreamPart<TOOLS>, TextStreamPart<TOOLS>>
--- a/packages/aiCore/src/core/runtime/index.ts
+++ b/packages/aiCore/src/core/runtime/index.ts
@ -16,7 +16,7 @@ export type {

 // === 便捷工厂函数 ===

-import { LanguageModelV1Middleware } from 'ai'
+import { LanguageModelV2Middleware } from '@ai-sdk/provider'

 import { type ProviderId, type ProviderSettingsMap } from '../../types'
 import { type AiPlugin } from '../plugins'
@ -54,7 +54,7 @@ export async function streamText<T extends ProviderId>(
  modelId: string,
  params: Parameters<RuntimeExecutor<T>['streamText']>[1],
  plugins?: AiPlugin[],
-  middlewares?: LanguageModelV1Middleware[]
+  middlewares?: LanguageModelV2Middleware[]
 ): Promise<ReturnType<RuntimeExecutor<T>['streamText']>> {
  const executor = createExecutor(providerId, options, plugins)
  return executor.streamText(modelId, params, { middlewares })
@ -69,7 +69,7 @@ export async function generateText<T extends ProviderId>(
  modelId: string,
  params: Parameters<RuntimeExecutor<T>['generateText']>[1],
  plugins?: AiPlugin[],
-  middlewares?: LanguageModelV1Middleware[]
+  middlewares?: LanguageModelV2Middleware[]
 ): Promise<ReturnType<RuntimeExecutor<T>['generateText']>> {
  const executor = createExecutor(providerId, options, plugins)
  return executor.generateText(modelId, params, { middlewares })
@ -84,7 +84,7 @@ export async function generateObject<T extends ProviderId>(
  modelId: string,
  params: Parameters<RuntimeExecutor<T>['generateObject']>[1],
  plugins?: AiPlugin[],
-  middlewares?: LanguageModelV1Middleware[]
+  middlewares?: LanguageModelV2Middleware[]
 ): Promise<ReturnType<RuntimeExecutor<T>['generateObject']>> {
  const executor = createExecutor(providerId, options, plugins)
  return executor.generateObject(modelId, params, { middlewares })
@ -99,7 +99,7 @@ export async function streamObject<T extends ProviderId>(
  modelId: string,
  params: Parameters<RuntimeExecutor<T>['streamObject']>[1],
  plugins?: AiPlugin[],
-  middlewares?: LanguageModelV1Middleware[]
+  middlewares?: LanguageModelV2Middleware[]
 ): Promise<ReturnType<RuntimeExecutor<T>['streamObject']>> {
  const executor = createExecutor(providerId, options, plugins)
  return executor.streamObject(modelId, params, { middlewares })
--- a/packages/aiCore/src/index.ts
+++ b/packages/aiCore/src/index.ts
@ -68,8 +68,10 @@ export type {
  TextStreamPart,
  // 工具相关类型
  Tool,
+  ToolCallUnion,
  ToolModelMessage,
  ToolResultPart,
+  ToolSet,
  UserModelMessage
 } from 'ai'
 export { defaultSettingsMiddleware, extractReasoningMiddleware, simulateStreamingMiddleware, smoothStream } from 'ai'
--- a/src/renderer/src/aiCore/AiSdkToChunkAdapter.ts
+++ b/src/renderer/src/aiCore/AiSdkToChunkAdapter.ts
@ -3,7 +3,8 @@
 * 用于将 AI SDK 的 fullStream 转换为 Cherry Studio 的 chunk 格式
 */

-import { TextStreamPart } from '@cherrystudio/ai-core'
+import { TextStreamPart, ToolSet } from '@cherrystudio/ai-core'
+import { MCPTool, WebSearchSource } from '@renderer/types'
 import { Chunk, ChunkType } from '@renderer/types/chunk'

 import { ToolCallChunkHandler } from './chunk/handleTooCallChunk'
@ -24,8 +25,11 @@ export interface CherryStudioChunk {
 */
 export class AiSdkToChunkAdapter {
  toolCallHandler: ToolCallChunkHandler
-  constructor(private onChunk: (chunk: Chunk) => void) {
-    this.toolCallHandler = new ToolCallChunkHandler(onChunk)
+  constructor(
+    private onChunk: (chunk: Chunk) => void,
+    private mcpTools: MCPTool[] = []
+  ) {
+    this.toolCallHandler = new ToolCallChunkHandler(onChunk, mcpTools)
  }

  /**
@ -47,7 +51,7 @@ export class AiSdkToChunkAdapter {
   * 读取 fullStream 并转换为 Cherry Studio chunks
   * @param fullStream AI SDK 的 fullStream (ReadableStream)
   */
-  private async readFullStream(fullStream: ReadableStream<TextStreamPart<any>>) {
+  private async readFullStream(fullStream: ReadableStream<TextStreamPart<ToolSet>>) {
    const reader = fullStream.getReader()
    const final = {
      text: '',
@ -73,84 +77,39 @@ export class AiSdkToChunkAdapter {
   * 转换 AI SDK chunk 为 Cherry Studio chunk 并调用回调
   * @param chunk AI SDK 的 chunk 数据
   */
-  private convertAndEmitChunk(chunk: any, final: { text: string; reasoning_content: string }) {
+  private convertAndEmitChunk(chunk: TextStreamPart<any>, final: { text: string; reasoning_content: string }) {
    console.log('AI SDK chunk type:', chunk.type, chunk)
    switch (chunk.type) {
      // === 文本相关事件 ===
-      case 'text-delta':
-        final.text += chunk.textDelta || ''
+      case 'text':
+        final.text += chunk.text || ''
        this.onChunk({
          type: ChunkType.TEXT_DELTA,
-          text: chunk.textDelta || ''
+          text: chunk.text || ''
+        })
+        break
+      case 'text-end':
+        this.onChunk({
+          type: ChunkType.TEXT_COMPLETE,
+          text: final.text || ''
        })
        break
-
      case 'reasoning':
        this.onChunk({
          type: ChunkType.THINKING_DELTA,
-          text: chunk.textDelta || '',
-          // 自定义字段
-          thinking_millsec: chunk.thinking_millsec || 0
+          text: chunk.text || '',
+          thinking_millsec: (chunk.providerMetadata?.metadata?.thinking_millsec as number) || 0
        })
        break
-      case 'redacted-reasoning':
-        this.onChunk({
-          type: ChunkType.THINKING_DELTA,
-          text: chunk.data || ''
-        })
-        break
-      case 'reasoning-signature':
+      case 'reasoning-end':
        this.onChunk({
          type: ChunkType.THINKING_COMPLETE,
-          text: chunk.text || '',
-          thinking_millsec: chunk.thinking_millsec || 0
+          text: (chunk.providerMetadata?.metadata?.thinking_content as string) || '',
+          thinking_millsec: (chunk.providerMetadata?.metadata?.thinking_millsec as number) || 0
        })
        break

      // === 工具调用相关事件（原始 AI SDK 事件，如果没有被中间件处理） ===
-      case 'tool-call-streaming-start':
-        // 开始流式工具调用
-        this.onChunk({
-          type: ChunkType.MCP_TOOL_CREATED,
-          tool_calls: [
-            {
-              id: chunk.toolCallId,
-              name: chunk.toolName,
-              args: {}
-            }
-          ]
-        })
-        break
-
-      case 'tool-call-delta':
-        // 工具调用参数的增量更新
-        this.onChunk({
-          type: ChunkType.MCP_TOOL_IN_PROGRESS,
-          responses: [
-            {
-              id: chunk.toolCallId,
-              tool: {
-                id: chunk.toolName,
-                // TODO: serverId,serverName
-                serverId: 'ai-sdk',
-                serverName: 'AI SDK',
-                name: chunk.toolName,
-                description: '',
-                inputSchema: {
-                  type: 'object',
-                  title: chunk.toolName,
-                  properties: {}
-                }
-              },
-              arguments: {},
-              status: 'invoking',
-              response: chunk.argsTextDelta,
-              toolCallId: chunk.toolCallId
-            }
-          ]
-        })
-        break
-
      case 'tool-call':
        // 原始的工具调用（未被中间件处理）
        this.toolCallHandler.handleToolCall(chunk)
@ -160,6 +119,11 @@ export class AiSdkToChunkAdapter {
        // 原始的工具调用结果（未被中间件处理）
        this.toolCallHandler.handleToolResult(chunk)
        break
+      // case 'start':
+      //   this.onChunk({
+      //     type: ChunkType.LLM_RESPONSE_CREATED
+      //   })
+      //   break

      // === 步骤相关事件 ===
      // TODO: 需要区分接口开始和步骤开始
@ -168,13 +132,17 @@ export class AiSdkToChunkAdapter {
      //     type: ChunkType.LLM_RESPONSE_CREATED
      //   })
      //   break
-      case 'step-finish':
-        this.onChunk({
-          type: ChunkType.TEXT_COMPLETE,
-          text: final.text || '' // TEXT_COMPLETE 需要 text 字段
-        })
-        final.text = ''
-        break
+      // case 'step-finish':
+      //   this.onChunk({
+      //     type: ChunkType.TEXT_COMPLETE,
+      //     text: final.text || '' // TEXT_COMPLETE 需要 text 字段
+      //   })
+      //   final.text = ''
+      //   break
+
+      // case 'finish-step': {
+      //   const { totalUsage, finishReason, providerMetadata } = chunk
+      // }

      case 'finish':
        this.onChunk({
@ -183,13 +151,13 @@ export class AiSdkToChunkAdapter {
            text: final.text || '',
            reasoning_content: final.reasoning_content || '',
            usage: {
-              completion_tokens: chunk.usage.completionTokens || 0,
-              prompt_tokens: chunk.usage.promptTokens || 0,
-              total_tokens: chunk.usage.totalTokens || 0
+              completion_tokens: chunk.totalUsage.outputTokens || 0,
+              prompt_tokens: chunk.totalUsage.inputTokens || 0,
+              total_tokens: chunk.totalUsage.totalTokens || 0
            },
-            metrics: chunk.usage
+            metrics: chunk.totalUsage
              ? {
-                  completion_tokens: chunk.usage.completionTokens || 0,
+                  completion_tokens: chunk.totalUsage.outputTokens || 0,
                  time_completion_millsec: 0
                }
              : undefined
@ -201,13 +169,13 @@ export class AiSdkToChunkAdapter {
            text: final.text || '',
            reasoning_content: final.reasoning_content || '',
            usage: {
-              completion_tokens: chunk.usage.completionTokens || 0,
-              prompt_tokens: chunk.usage.promptTokens || 0,
-              total_tokens: chunk.usage.totalTokens || 0
+              completion_tokens: chunk.totalUsage.outputTokens || 0,
+              prompt_tokens: chunk.totalUsage.inputTokens || 0,
+              total_tokens: chunk.totalUsage.totalTokens || 0
            },
-            metrics: chunk.usage
+            metrics: chunk.totalUsage
              ? {
-                  completion_tokens: chunk.usage.completionTokens || 0,
+                  completion_tokens: chunk.totalUsage.outputTokens || 0,
                  time_completion_millsec: 0
                }
              : undefined
@ -217,30 +185,24 @@ export class AiSdkToChunkAdapter {

      // === 源和文件相关事件 ===
      case 'source':
-        // 源信息，可以映射到知识搜索完成
        this.onChunk({
-          type: ChunkType.KNOWLEDGE_SEARCH_COMPLETE,
-          knowledge: [
-            {
-              id: Number(chunk.source.id) || Date.now(),
-              content: chunk.source.title || '',
-              sourceUrl: chunk.source.url || '',
-              type: 'url'
-            }
-          ]
-        })
-        break
-
-      case 'file':
-        // 文件相关事件，可能是图片生成
-        this.onChunk({
-          type: ChunkType.IMAGE_COMPLETE,
-          image: {
-            type: 'base64',
-            images: [chunk.base64]
+          type: ChunkType.LLM_WEB_SEARCH_COMPLETE,
+          llm_web_search: {
+            source: WebSearchSource.AISDK,
+            results: [{}]
          }
        })
        break
+      // case 'file':
+      //   // 文件相关事件，可能是图片生成
+      //   this.onChunk({
+      //     type: ChunkType.IMAGE_COMPLETE,
+      //     image: {
+      //       type: 'base64',
+      //       images: [chunk.base64]
+      //     }
+      //   })
+      //   break
      case 'error':
        this.onChunk({
          type: ChunkType.ERROR,
--- a/src/renderer/src/aiCore/chunk/handleTooCallChunk.ts
+++ b/src/renderer/src/aiCore/chunk/handleTooCallChunk.ts
@ -4,8 +4,9 @@
 * 提供工具调用相关的处理API，每个交互使用一个新的实例
 */

+import { ToolCallUnion, ToolSet } from '@cherrystudio/ai-core/index'
 import Logger from '@renderer/config/logger'
-import { MCPToolResponse } from '@renderer/types'
+import { MCPTool, MCPToolResponse } from '@renderer/types'
 import { Chunk, ChunkType } from '@renderer/types/chunk'

 /**
@ -19,10 +20,13 @@ export class ToolCallChunkHandler {
      toolCallId: string
      toolName: string
      args: any
-      //   mcpTool: MCPTool
+      mcpTool: MCPTool
    }
  >()
-  constructor(private onChunk: (chunk: Chunk) => void) {}
+  constructor(
+    private onChunk: (chunk: Chunk) => void,
+    private mcpTools: MCPTool[]
+  ) {}

  //   /**
  //    * 设置 onChunk 回调
@ -34,10 +38,14 @@ export class ToolCallChunkHandler {
  /**
   * 处理工具调用事件
   */
-  public handleToolCall(chunk: any): void {
+  public handleToolCall(
+    chunk: {
+      type: 'tool-call'
+    } & ToolCallUnion<ToolSet>
+  ): void {
    const toolCallId = chunk.toolCallId
    const toolName = chunk.toolName
-    const args = chunk.args || {}
+    const args = chunk.input || {}

    if (!toolCallId || !toolName) {
      Logger.warn(`🔧 [ToolCallChunkHandler] Invalid tool call chunk: missing toolCallId or toolName`)
@ -51,17 +59,14 @@ export class ToolCallChunkHandler {
    this.activeToolCalls.set(toolCallId, {
      toolCallId,
      toolName,
-      args
-      //   mcpTool
+      args,
+      mcpTool: this.mcpTools.find((tool) => tool.name === toolName)!
    })

    // 创建 MCPToolResponse 格式
    const toolResponse: MCPToolResponse = {
      id: toolCallId,
-      tool: {
-        id: toolCallId,
-        name: toolName
-      },
+      tool: this.activeToolCalls.get(toolCallId)!.mcpTool,
      arguments: args,
      status: 'invoking',
      toolCallId: toolCallId
@ -98,10 +103,7 @@ export class ToolCallChunkHandler {
    // 创建工具调用结果的 MCPToolResponse 格式
    const toolResponse: MCPToolResponse = {
      id: toolCallId,
-      tool: {
-        id: toolCallId,
-        name: toolCallInfo.toolName
-      },
+      tool: toolCallInfo.mcpTool,
      arguments: toolCallInfo.args,
      status: 'done',
      response: {
--- a/src/renderer/src/aiCore/index_new.ts
+++ b/src/renderer/src/aiCore/index_new.ts
@ -195,7 +195,7 @@ export default class ModernAiProvider {
    // 创建带有中间件的执行器
    if (middlewareConfig.onChunk) {
      // 流式处理 - 使用适配器
-      const adapter = new AiSdkToChunkAdapter(middlewareConfig.onChunk)
+      const adapter = new AiSdkToChunkAdapter(middlewareConfig.onChunk, middlewareConfig.mcpTools)
      console.log('最终params', params)
      const streamResult = await executor.streamText(
        modelId,
--- a/src/renderer/src/aiCore/middleware/aisdk/AiSdkMiddlewareBuilder.ts
+++ b/src/renderer/src/aiCore/middleware/aisdk/AiSdkMiddlewareBuilder.ts
@ -1,8 +1,4 @@
-import {
-  extractReasoningMiddleware,
-  LanguageModelV1Middleware,
-  simulateStreamingMiddleware
-} from '@cherrystudio/ai-core'
+import { LanguageModelV2Middleware, simulateStreamingMiddleware } from '@cherrystudio/ai-core'
 import type { MCPTool, Model, Provider } from '@renderer/types'
 import type { Chunk } from '@renderer/types/chunk'

@ -26,7 +22,7 @@ export interface AiSdkMiddlewareConfig {
 */
 export interface NamedAiSdkMiddleware {
  name: string
-  middleware: LanguageModelV1Middleware
+  middleware: LanguageModelV2Middleware
 }

 /**
@ -75,7 +71,7 @@ export class AiSdkMiddlewareBuilder {
  /**
   * 构建最终的中间件数组
   */
-  public build(): LanguageModelV1Middleware[] {
+  public build(): LanguageModelV2Middleware[] {
    return this.middlewares.map((m) => m.middleware)
  }

@ -106,7 +102,7 @@ export class AiSdkMiddlewareBuilder {
 * 根据配置构建AI SDK中间件的工厂函数
 * 这里要注意构建顺序，因为有些中间件需要依赖其他中间件的结果
 */
-export function buildAiSdkMiddlewares(config: AiSdkMiddlewareConfig): LanguageModelV1Middleware[] {
+export function buildAiSdkMiddlewares(config: AiSdkMiddlewareConfig): LanguageModelV2Middleware[] {
  const builder = new AiSdkMiddlewareBuilder()

  // 1. 根据provider添加特定中间件
@ -143,10 +139,10 @@ function addProviderSpecificMiddlewares(builder: AiSdkMiddlewareBuilder, config:
      // Anthropic特定中间件
      break
    case 'openai':
-      builder.add({
-        name: 'thinking-tag-extraction',
-        middleware: extractReasoningMiddleware({ tagName: 'think' })
-      })
+      // builder.add({
+      //   name: 'thinking-tag-extraction',
+      //   middleware: extractReasoningMiddleware({ tagName: 'think' })
+      // })
      break
    case 'gemini':
      // Gemini特定中间件
--- a/src/renderer/src/aiCore/plugins/reasoningTimePlugin.ts
+++ b/src/renderer/src/aiCore/plugins/reasoningTimePlugin.ts
@ -1,4 +1,4 @@
-import { definePlugin } from '@cherrystudio/ai-core'
+import { definePlugin, TextStreamPart, ToolSet } from '@cherrystudio/ai-core'

 export default definePlugin({
  name: 'reasoningTimePlugin',
@ -8,57 +8,62 @@ export default definePlugin({
    let thinkingStartTime = 0
    let hasStartedThinking = false
    let accumulatedThinkingContent = ''
+    let reasoningBlockId = ''

-    return new TransformStream({
-      transform(chunk, controller) {
-        if (chunk.type !== 'reasoning') {
-          // === 处理 reasoning 结束  ===
-          if (hasStartedThinking) {
-            console.log(`[ReasoningPlugin] Ending reasoning.`)
-
-            // 生成 reasoning-signature
-            controller.enqueue({
-              type: 'reasoning-signature',
-              text: accumulatedThinkingContent,
-              thinking_millsec: performance.now() - thinkingStartTime
-            })
-
-            // 重置状态
-            accumulatedThinkingContent = ''
-            hasStartedThinking = false
-            thinkingStartTime = 0
-          }
-
-          controller.enqueue(chunk)
-          return
-        }
-
+    return new TransformStream<TextStreamPart<ToolSet>, TextStreamPart<ToolSet>>({
+      transform(chunk: TextStreamPart<ToolSet>, controller: TransformStreamDefaultController<TextStreamPart<ToolSet>>) {
        // === 处理 reasoning 类型 ===
+        if (chunk.type === 'reasoning') {
+          if (!hasStartedThinking) {
+            hasStartedThinking = true
+            thinkingStartTime = performance.now()
+            reasoningBlockId = chunk.id
+          }
+          accumulatedThinkingContent += chunk.text

-        // 1. 时间跟踪逻辑
-        if (!hasStartedThinking) {
-          hasStartedThinking = true
-          thinkingStartTime = performance.now()
-          console.log(`[ReasoningPlugin] Starting reasoning session`)
+          controller.enqueue({
+            ...chunk,
+            providerMetadata: {
+              ...chunk.providerMetadata,
+              metadata: {
+                ...chunk.providerMetadata?.metadata,
+                thinking_millsec: performance.now() - thinkingStartTime,
+                thinking_content: accumulatedThinkingContent
+              }
+            }
+          })
+        } else if (hasStartedThinking) {
+          controller.enqueue({
+            type: 'reasoning-end',
+            id: reasoningBlockId,
+            providerMetadata: {
+              metadata: {
+                thinking_millsec: performance.now() - thinkingStartTime,
+                thinking_content: accumulatedThinkingContent
+              }
+            }
+          })
+          accumulatedThinkingContent = ''
+          hasStartedThinking = false
+          thinkingStartTime = 0
+          reasoningBlockId = ''
+          controller.enqueue(chunk)
+        } else {
+          controller.enqueue(chunk)
        }
-        accumulatedThinkingContent += chunk.textDelta
-
-        // 2. 直接透传 chunk，并附加上时间
-        console.log(`[ReasoningPlugin] Forwarding reasoning chunk: "${chunk.textDelta}"`)
-        controller.enqueue({
-          ...chunk,
-          thinking_millsec: performance.now() - thinkingStartTime
-        })
      },

-      // === flush 处理流结束时仍在reasoning状态的场景 ===
      flush(controller) {
        if (hasStartedThinking) {
-          console.log(`[ReasoningPlugin] Final flush for reasoning-signature.`)
          controller.enqueue({
-            type: 'reasoning-signature',
-            text: accumulatedThinkingContent,
-            thinking_millsec: performance.now() - thinkingStartTime
+            type: 'reasoning-end',
+            id: reasoningBlockId,
+            providerMetadata: {
+              metadata: {
+                thinking_millsec: performance.now() - thinkingStartTime,
+                thinking_content: accumulatedThinkingContent
+              }
+            }
          })
        }
      }
--- a/src/renderer/src/aiCore/utils/mcp.ts
+++ b/src/renderer/src/aiCore/utils/mcp.ts
@ -48,7 +48,7 @@ export function convertMcpToolsToAiSdkTools(mcpTools: MCPTool[]): Record<string,
    console.log('mcpTool', mcpTool.inputSchema)
    tools[mcpTool.name] = aiSdk.tool<any, ToolCallResult>({
      description: mcpTool.description || `Tool from ${mcpTool.serverName}`,
-      parameters: aiSdk.jsonSchema(mcpTool.inputSchema as JSONSchema7),
+      inputSchema: aiSdk.jsonSchema(mcpTool.inputSchema as JSONSchema7),
      execute: async (params): Promise<ToolCallResult> => {
        console.log('execute_params', params)
        // 创建适配的 MCPToolResponse 对象
--- a/src/renderer/src/types/index.ts
+++ b/src/renderer/src/types/index.ts
@ -545,7 +545,8 @@ export enum WebSearchSource {
  QWEN = 'qwen',
  HUNYUAN = 'hunyuan',
  ZHIPU = 'zhipu',
-  GROK = 'grok'
+  GROK = 'grok',
+  AISDK = 'ai-sdk'
 }

 export type WebSearchResponse = {
--- a/yarn.lock
+++ b/yarn.lock
@ -956,6 +956,7 @@ __metadata:
    ollama-ai-provider: "npm:^1.2.0"
    tsdown: "npm:^0.12.9"
    typescript: "npm:^5.0.0"
+    zod: "npm:^3.25.0"
  peerDependenciesMeta:
    "@ai-sdk/amazon-bedrock":
      optional: true
@ -20140,6 +20141,13 @@ __metadata:
  languageName: node
  linkType: hard

+"zod@npm:^3.25.0":
+  version: 3.25.74
+  resolution: "zod@npm:3.25.74"
+  checksum: 10c0/59e38b046ac333b5bd1ba325a83b6798721227cbfb1e69dfc7159bd7824b904241ab923026edb714fafefec3624265ae374a70aee9a5a45b365bd31781ffa105
+  languageName: node
+  linkType: hard
+
 "zustand@npm:^4.4.0":
  version: 4.5.6
  resolution: "zustand@npm:4.5.6"