diff --git a/.github/workflows/deploy.yml b/.github/workflows/deploy.yml
index 5bfbabd..2d960bb 100644
--- a/.github/workflows/deploy.yml
+++ b/.github/workflows/deploy.yml
@@ -34,7 +34,9 @@ jobs:
       - name: Write .env file
         run: |
           cat > /tmp/.env << EOF
-          OPENAI_API_KEY=${{ secrets.OPENAI_API_KEY }}
+          AI_API_KEY=${{ secrets.AI_API_KEY }}
+          AI_MODEL=${{ vars.AI_MODEL }}
+          AI_BASE_URL=${{ vars.AI_BASE_URL }}
           TENCENT_SECRET_ID=${{ secrets.TENCENT_SECRET_ID }}
           TENCENT_SECRET_KEY=${{ secrets.TENCENT_SECRET_KEY }}
           EOF
diff --git a/backend/.claude/settings.local.json b/backend/.claude/settings.local.json
index f8a396f..5209da3 100644
--- a/backend/.claude/settings.local.json
+++ b/backend/.claude/settings.local.json
@@ -16,7 +16,9 @@
       "Bash(fi)",
       "Bash(done)",
       "Bash(npm run format:*)",
-      "Bash(timeout 8 npm run dev:*)"
+      "Bash(timeout 8 npm run dev:*)",
+      "Bash(git -C \"C:\\\\Users\\\\Shijia Huang\\\\WeChatProjects\\\\miniprogram-2\" diff HEAD -- backend/src/constants/config.ts)",
+      "Bash(git -C \"C:\\\\Users\\\\Shijia Huang\\\\WeChatProjects\\\\miniprogram-2\" diff HEAD -- backend/src/clients/openai.client.ts)"
     ]
   }
 }
diff --git a/backend/.env.example b/backend/.env.example
index b6fd2ca..cecd011 100644
--- a/backend/.env.example
+++ b/backend/.env.example
@@ -27,9 +27,10 @@ GAME_MOVE_TIMEOUT=30000
 # Logging
 LOG_LEVEL=debug
 
-# OpenAI Configuration
-OPENAI_API_KEY=YOUR_OPENAI_API_KEY
-OPENAI_MODEL=gpt-4o
+# AI Configuration
+AI_API_KEY=YOUR_AI_API_KEY
+AI_MODEL=deepseek-chat
+AI_BASE_URL=https://api.deepseek.com/v1
 
 # LLM Worker Configuration
 LLM_WORKER_POLL_INTERVAL_MS=500
diff --git a/backend/src/clients/openai.client.ts b/backend/src/clients/openai.client.ts
index c0cb98d..e16b22a 100644
--- a/backend/src/clients/openai.client.ts
+++ b/backend/src/clients/openai.client.ts
@@ -12,7 +12,7 @@ import crypto from 'crypto';
 
 import OpenAI from 'openai';
 
-import { OPENAI_CONFIG } from '../constants/config';
+import { AI_CONFIG } from '../constants/config';
 import {
     JUDGMENT_SYSTEM_PROMPT,
     buildJudgmentUserContent,
@@ -30,13 +30,13 @@ const REQUEST_TIMEOUT_MS = 60_000;
  * Create a singleton OpenAI client instance
  */
 function createClient(): OpenAI {
-    if (!OPENAI_CONFIG.API_KEY) {
-        throw new Error('OPENAI_API_KEY 未配置，无法调用 LLM');
+    if (!AI_CONFIG.API_KEY) {
+        throw new Error('AI_API_KEY 未配置，无法调用 LLM');
     }
 
     return new OpenAI({
-        apiKey: OPENAI_CONFIG.API_KEY,
-        ...(OPENAI_CONFIG.BASE_URL ? { baseURL: OPENAI_CONFIG.BASE_URL } : {}),
+        apiKey: AI_CONFIG.API_KEY,
+        ...(AI_CONFIG.BASE_URL ? { baseURL: AI_CONFIG.BASE_URL } : {}),
         timeout: REQUEST_TIMEOUT_MS,
     });
 }
@@ -267,7 +267,7 @@ export async function createJudgmentVerdict(
 
     const response = await client.chat.completions.create(
         {
-            model: OPENAI_CONFIG.MODEL,
+            model: AI_CONFIG.MODEL,
             messages: [
                 {
                     role: 'system',
diff --git a/backend/src/constants/config.ts b/backend/src/constants/config.ts
index 8b0cd4c..3f6113c 100644
--- a/backend/src/constants/config.ts
+++ b/backend/src/constants/config.ts
@@ -37,13 +37,13 @@ export const VERDICT_CONFIG = {
     MAX_RETRIES: 3,
 } as const;
 
-export const OPENAI_CONFIG = {
-    /** OpenAI API key */
-    API_KEY: process.env.OPENAI_API_KEY || '',
-    /** OpenAI model to use */
-    MODEL: process.env.OPENAI_MODEL || 'gpt-4o',
-    /** Optional base URL for OpenAI-compatible APIs */
-    BASE_URL: process.env.OPENAI_BASE_URL,
+export const AI_CONFIG = {
+    /** AI API key */
+    API_KEY: process.env.AI_API_KEY || '',
+    /** AI model to use */
+    MODEL: process.env.AI_MODEL || '',
+    /** Optional base URL for AI-compatible APIs */
+    BASE_URL: process.env.AI_BASE_URL || '',
 } as const;
 
 export const TENCENT_CONFIG = {