refacto: reasoning cache implementation and update import paths

src/main/apiServer/adapters/AiSdkToAnthropicSSE.ts

@@ -38,7 +38,7 @@ import type {
 import { loggerService } from '@logger'
 import { type FinishReason, type LanguageModelUsage, type TextStreamPart, type ToolSet } from 'ai'
 
-import { googleReasoningCache, openRouterReasoningCache } from '../../services/CacheService'
+import { googleReasoningCache, openRouterReasoningCache } from '../services/reasoning-cache'
 
 const logger = loggerService.withContext('AiSdkToAnthropicSSE')
 

src/main/apiServer/services/reasoning-cache.ts

@@ -0,0 +1,45 @@
+/**
+ * Reasoning Cache Service
+ *
+ * Manages reasoning-related caching for AI providers that support thinking/reasoning modes.
+ * This includes Google Gemini's thought signatures and OpenRouter's reasoning details.
+ */
+
+import type { ReasoningDetailUnion } from '@main/apiServer/adapters/openrouter'
+import { CacheService } from '@main/services/CacheService'
+
+/**
+ * Interface for reasoning cache
+ */
+export interface IReasoningCache<T> {
+  set(key: string, value: T): void
+  get(key: string): T | undefined
+}
+
+/**
+ * Cache duration: 30 minutes
+ * Reasoning data is typically only needed within a short conversation context
+ */
+const REASONING_CACHE_DURATION = 30 * 60 * 1000
+
+/**
+ * Google Gemini reasoning cache
+ *
+ * Stores thought signatures for Gemini 3 models to handle multi-turn conversations
+ * where the model needs to maintain thinking context across tool calls.
+ */
+export const googleReasoningCache: IReasoningCache<string> = {
+  set: (key, value) => CacheService.set(`google-reasoning:${key}`, value, REASONING_CACHE_DURATION),
+  get: (key) => CacheService.get(`google-reasoning:${key}`) || undefined
+}
+
+/**
+ * OpenRouter reasoning cache
+ *
+ * Stores reasoning details from OpenRouter responses to preserve thinking tokens
+ * and reasoning metadata across the conversation flow.
+ */
+export const openRouterReasoningCache: IReasoningCache<ReasoningDetailUnion[]> = {
+  set: (key, value) => CacheService.set(`openrouter-reasoning:${key}`, value, REASONING_CACHE_DURATION),
+  get: (key) => CacheService.get(`openrouter-reasoning:${key}`) || undefined
+}

src/main/apiServer/services/unified-messages.ts

@@ -42,7 +42,7 @@ import { net } from 'electron'
 import type { Response } from 'express'
 import * as z from 'zod'
 
-import { googleReasoningCache, openRouterReasoningCache } from '../../services/CacheService'
+import { googleReasoningCache, openRouterReasoningCache } from './reasoning-cache'
 
 const logger = loggerService.withContext('UnifiedMessagesService')
 

src/main/services/CacheService.ts

@@ -1,19 +1,9 @@
-import type { ReasoningDetailUnion } from '@main/apiServer/adapters/openrouter'
-
 interface CacheItem<T> {
   data: T
   timestamp: number
   duration: number
 }
 
-/**
- * Interface for reasoning cache
- */
-export interface IReasoningCache<T> {
-  set(key: string, value: T): void
-  get(key: string): T | undefined
-}
-
 export class CacheService {
   private static cache: Map<string, CacheItem<any>> = new Map()
 
@@ -82,14 +72,3 @@ export class CacheService {
     return true
   }
 }
-
-// Singleton cache instances using CacheService
-export const googleReasoningCache: IReasoningCache<string> = {
-  set: (key, value) => CacheService.set(`google-reasoning:${key}`, value, 30 * 60 * 1000),
-  get: (key) => CacheService.get(`google-reasoning:${key}`) || undefined
-}
-
-export const openRouterReasoningCache: IReasoningCache<ReasoningDetailUnion[]> = {
-  set: (key, value) => CacheService.set(`openrouter-reasoning:${key}`, value, 30 * 60 * 1000),
-  get: (key) => CacheService.get(`openrouter-reasoning:${key}`) || undefined
-}