🐛 fix: add SSRF protection (#10152)

.env.example

@@ -13,6 +13,17 @@
 # Default is '0' (enabled)
 # ENABLED_CSP=1
 
+# SSRF Protection Settings
+# Set to '1' to allow connections to private IP addresses (disable SSRF protection)
+# WARNING: Only enable this in trusted environments
+# Default is '0' (SSRF protection enabled)
+# SSRF_ALLOW_PRIVATE_IP_ADDRESS=0
+
+# Whitelist of allowed private IP addresses (comma-separated)
+# Only takes effect when SSRF_ALLOW_PRIVATE_IP_ADDRESS is '0'
+# Example: Allow specific internal servers while keeping SSRF protection
+# SSRF_ALLOW_IP_ADDRESS_LIST=192.168.1.100,10.0.0.50
+
 ########################################
 ########## AI Provider Service #########
 ########################################

docs/self-hosting/environment-variables/basic.mdx

@@ -127,16 +127,62 @@ For specific content, please refer to the [Feature Flags](/docs/self-hosting/adv
 ### `SSRF_ALLOW_PRIVATE_IP_ADDRESS`
 
 - Type: Optional
-- Description: Allow to connect private IP address. In a trusted environment, it can be set to true to turn off SSRF protection.
+- Description: Controls whether to allow connections to private IP addresses. Set to `1` to disable SSRF protection and allow all private IP addresses. In a trusted environment (e.g., internal network), this can be enabled to allow access to internal resources.
 - Default: `0`
 - Example: `1` or `0`
 
+<Callout type="warning">
+  **Security Notice**: Enabling this option will disable SSRF protection and allow connections to private
+  IP addresses (127.0.0.0/8, 10.0.0.0/8, 172.16.0.0/12, 192.168.0.0/16, etc.). Only enable this in
+  trusted environments where you need to access internal network resources.
+</Callout>
+
+**Use Cases**:
+
+LobeChat performs SSRF security checks in the following scenarios:
+
+1. **Image/Video URL to Base64 Conversion**: When processing media messages (e.g., vision models, multimodal models), LobeChat converts image and video URLs to base64 format. This check prevents malicious users from accessing internal network resources.
+
+   Examples:
+
+   - Image: A user sends an image message with URL `http://192.168.1.100/admin/secrets.png`
+   - Video: A user sends a video message with URL `http://10.0.0.50/internal/meeting.mp4`
+
+   Without SSRF protection, these requests could expose internal network resources.
+
+2. **Web Crawler**: When using web crawling features to fetch external content.
+
+3. **Proxy Requests**: When proxying external API requests.
+
+**Configuration Examples**:
+
+```bash
+# Scenario 1: Public deployment (recommended)
+# Block all private IP addresses for security
+SSRF_ALLOW_PRIVATE_IP_ADDRESS=0
+
+# Scenario 2: Internal deployment
+# Allow all private IP addresses to access internal image servers
+SSRF_ALLOW_PRIVATE_IP_ADDRESS=1
+
+# Scenario 3: Hybrid deployment (most common)
+# Block private IPs by default, but allow specific trusted internal servers
+SSRF_ALLOW_PRIVATE_IP_ADDRESS=0
+SSRF_ALLOW_IP_ADDRESS_LIST=192.168.1.100,10.0.0.50
+```
+
 ### `SSRF_ALLOW_IP_ADDRESS_LIST`
 
 - Type: Optional
-- Description: Allow private IP address list, multiple IP addresses are separated by commas. Only when `SSRF_ALLOW_PRIVATE_IP_ADDRESS` is `0`, it takes effect.
+- Description: Whitelist of allowed IP addresses, separated by commas. Only takes effect when `SSRF_ALLOW_PRIVATE_IP_ADDRESS` is `0`. Use this to allow specific internal IP addresses while keeping SSRF protection enabled for other private IPs.
 - Default: -
-- Example: `198.18.1.62,224.0.0.3`
+- Example: `192.168.1.100,10.0.0.50,172.16.0.10`
+
+**Common Use Cases**:
+
+- Allow access to internal image storage server: `192.168.1.100`
+- Allow access to internal API gateway: `10.0.0.50`
+- Allow access to internal documentation server: `172.16.0.10`
 
 ### `ENABLE_AUTH_PROTECTION`
 

docs/self-hosting/environment-variables/basic.zh-CN.mdx

@@ -123,16 +123,61 @@ LobeChat 在部署时提供了一些额外的配置项，你可以使用环境
 ### `SSRF_ALLOW_PRIVATE_IP_ADDRESS`
 
 - 类型：可选
-- 描述：是否允许连接私有 IP 地址。在可信环境中可以设置为 true 来关闭 SSRF 防护。
+- 描述：控制是否允许连接私有 IP 地址。设置为 `1` 时将关闭 SSRF 防护并允许所有私有 IP 地址。在可信环境（如内网部署）中，可以启用此选项以访问内部资源。
 - 默认值：`0`
-- 示例：`1` or `0`
+- 示例：`1` 或 `0`
+
+<Callout type="warning">
+  **安全提示**：启用此选项将关闭 SSRF 防护，允许连接私有 IP 地址段（127.0.0.0/8、10.0.0.0/8、172.16.0.0/12、192.168.0.0/16
+  等）。仅在需要访问内网资源的可信环境中启用。
+</Callout>
+
+**应用场景**：
+
+LobeChat 会在以下场景执行 SSRF 安全检查：
+
+1. **图片 / 视频 URL 转 Base64**：在处理媒体消息时（例如视觉模型、多模态模型），LobeChat 会将图片和视频 URL 转换为 base64 格式。此检查可防止恶意用户通过媒体 URL 访问内网资源。
+
+   举例：
+
+   - 图片：用户发送图片消息，URL 为 `http://192.168.1.100/admin/secrets.png`
+   - 视频：用户发送视频消息，URL 为 `http://10.0.0.50/internal/meeting.mp4`
+
+   若无 SSRF 防护，这些请求可能导致内网资源泄露。
+
+2. **网页爬取**：使用网页爬取功能获取外部内容时。
+
+3. **代理请求**：代理外部 API 请求时。
+
+**配置示例**：
+
+```bash
+# 场景 1：公网部署（推荐）
+# 阻止所有私有 IP 访问，保证安全
+SSRF_ALLOW_PRIVATE_IP_ADDRESS=0
+
+# 场景 2：内网部署
+# 允许所有私有 IP，可访问内网图片服务器等资源
+SSRF_ALLOW_PRIVATE_IP_ADDRESS=1
+
+# 场景 3：混合部署（最常见）
+# 默认阻止私有 IP，但允许特定可信的内网服务器
+SSRF_ALLOW_PRIVATE_IP_ADDRESS=0
+SSRF_ALLOW_IP_ADDRESS_LIST=192.168.1.100,10.0.0.50
+```
 
 ### `SSRF_ALLOW_IP_ADDRESS_LIST`
 
 - 类型：可选
-- 说明：允许的私有 IP 地址列表，多个 IP 地址用逗号分隔。仅在 `SSRF_ALLOW_PRIVATE_IP_ADDRESS` 为 `0` 时生效。
+- 描述：允许访问的 IP 地址白名单，多个 IP 地址用逗号分隔。仅在 `SSRF_ALLOW_PRIVATE_IP_ADDRESS` 为 `0` 时生效。使用此选项可以在保持 SSRF 防护的同时，允许访问特定的内网 IP 地址。
 - 默认值：-
-- 示例：`198.18.1.62,224.0.0.3`
+- 示例：`192.168.1.100,10.0.0.50,172.16.0.10`
+
+**常见使用场景**：
+
+- 允许访问内网图片存储服务器：`192.168.1.100`
+- 允许访问内网 API 网关：`10.0.0.50`
+- 允许访问内网文档服务器：`172.16.0.10`
 
 ### `ENABLE_AUTH_PROTECTION`
 

package.json

@@ -154,6 +154,7 @@
     "@lobechat/database": "workspace:*",
     "@lobechat/electron-client-ipc": "workspace:*",
     "@lobechat/electron-server-ipc": "workspace:*",
+    "@lobechat/fetch-sse": "workspace:*",
     "@lobechat/file-loaders": "workspace:*",
     "@lobechat/model-runtime": "workspace:*",
     "@lobechat/observability-otel": "workspace:*",

packages/fetch-sse/package.json

@@ -0,0 +1,29 @@
+{
+  "name": "@lobechat/fetch-sse",
+  "version": "1.0.0",
+  "private": true,
+  "description": "SSE fetch utilities with streaming support",
+  "exports": {
+    ".": {
+      "types": "./src/index.ts",
+      "default": "./src/index.ts"
+    },
+    "./parseError": {
+      "types": "./src/parseError.ts",
+      "default": "./src/parseError.ts"
+    }
+  },
+  "main": "./src/index.ts",
+  "types": "./src/index.ts",
+  "scripts": {
+    "test": "vitest",
+    "test:coverage": "vitest --coverage --silent='passed-only'"
+  },
+  "dependencies": {
+    "@lobechat/const": "workspace:*",
+    "@lobechat/model-runtime": "workspace:*",
+    "@lobechat/types": "workspace:*",
+    "@lobechat/utils": "workspace:*",
+    "i18next": "^24.2.1"
+  }
+}

packages/fetch-sse/src/__tests__/fetchSSE.test.ts

@@ -1,10 +1,10 @@
 import { MESSAGE_CANCEL_FLAT } from '@lobechat/const';
 import { ChatMessageError } from '@lobechat/types';
+import { FetchEventSourceInit } from '@lobechat/utils/client/fetchEventSource/index';
+import { fetchEventSource } from '@lobechat/utils/client/fetchEventSource/index';
+import { sleep } from '@lobechat/utils/sleep';
 import { afterEach, describe, expect, it, vi } from 'vitest';
 
-import { FetchEventSourceInit } from '../../client/fetchEventSource';
-import { fetchEventSource } from '../../client/fetchEventSource';
-import { sleep } from '../../sleep';
 import { fetchSSE } from '../fetchSSE';
 
 // 模拟 i18next
@@ -12,7 +12,7 @@ vi.mock('i18next', () => ({
   t: vi.fn((key) => `translated_${key}`),
 }));
 
-vi.mock('../../client/fetchEventSource', () => ({
+vi.mock('@lobechat/utils/client/fetchEventSource/index', () => ({
   fetchEventSource: vi.fn(),
 }));
 

packages/fetch-sse/src/__tests__/parseError.test.ts

@@ -1,14 +1,14 @@
 import { ErrorResponse } from '@lobechat/types';
-import { afterEach, describe, expect, it, vi } from 'vitest';
+import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
 
 import { getMessageError } from '../parseError';
 
-// 模拟 i18next
+// Mock i18next
 vi.mock('i18next', () => ({
   t: vi.fn((key) => `translated_${key}`),
 }));
 
-// 模拟 Response
+// Mock Response
 const createMockResponse = (body: any, ok: boolean, status: number = 200) => ({
   ok,
   status,
@@ -38,11 +38,14 @@ const createMockResponse = (body: any, ok: boolean, status: number = 200) => ({
   },
 });
 
-// 在每次测试后清理所有模拟
 afterEach(() => {
   vi.restoreAllMocks();
 });
 
+beforeEach(() => {
+  vi.clearAllMocks();
+});
+
 describe('getMessageError', () => {
   it('should handle business error correctly', async () => {
     const mockErrorResponse: ErrorResponse = {

packages/fetch-sse/src/fetchSSE.ts

@@ -12,9 +12,9 @@ import {
   ResponseAnimation,
   ResponseAnimationStyle,
 } from '@lobechat/types';
+import { fetchEventSource } from '@lobechat/utils/client/fetchEventSource/index';
+import { nanoid } from '@lobechat/utils/uuid';
 
-import { fetchEventSource } from '../client/fetchEventSource';
-import { nanoid } from '../uuid';
 import { getMessageError } from './parseError';
 
 type SSEFinishType = 'done' | 'error' | 'abort';

packages/fetch-sse/src/parseError.ts

@@ -1,7 +1,7 @@
 import { ChatMessageError, ErrorResponse, ErrorType } from '@lobechat/types';
 import { t } from 'i18next';
 
-export const getMessageError = async (response: Response) => {
+export const getMessageError = async (response: Response): Promise<ChatMessageError> => {
   let chatMessageError: ChatMessageError;
 
   // try to get the biz error
@@ -9,13 +9,13 @@ export const getMessageError = async (response: Response) => {
     const data = (await response.json()) as ErrorResponse;
     chatMessageError = {
       body: data.body,
-      message: t(`response.${data.errorType}` as any, { ns: 'error' }),
+      message: t(`response.${data.errorType}`, { ns: 'error' }),
       type: data.errorType,
     };
   } catch {
     // if not return, then it's a common error
     chatMessageError = {
-      message: t(`response.${response.status}` as any, { ns: 'error' }),
+      message: t(`response.${response.status}`, { ns: 'error' }),
       type: response.status as ErrorType,
     };
   }

packages/model-runtime/src/core/contextBuilders/anthropic.test.ts

@@ -1,8 +1,8 @@
+import { imageUrlToBase64 } from '@lobechat/utils';
 import { OpenAI } from 'openai';
-import { describe, expect, it, vi } from 'vitest';
+import { beforeEach, describe, expect, it, vi } from 'vitest';
 
 import { OpenAIChatMessage, UserMessageContentPart } from '../../types/chat';
-import { imageUrlToBase64 } from '../../utils/imageToBase64';
 import { parseDataUri } from '../../utils/uriParser';
 import {
   buildAnthropicBlock,
@@ -12,16 +12,22 @@ import {
 } from './anthropic';
 
 // Mock the parseDataUri function since it's an implementation detail
-vi.mock('../../utils/uriParser', () => ({
-  parseDataUri: vi.fn().mockReturnValue({
-    mimeType: 'image/jpeg',
-    base64: 'base64EncodedString',
-    type: 'base64',
-  }),
+vi.mock('../../utils/uriParser');
+vi.mock('@lobechat/utils', () => ({
+  imageUrlToBase64: vi.fn(),
 }));
-vi.mock('../../utils/imageToBase64');
 
 describe('anthropicHelpers', () => {
+  beforeEach(() => {
+    vi.resetAllMocks();
+    // Set default mock implementation for parseDataUri
+    vi.mocked(parseDataUri).mockReturnValue({
+      mimeType: 'image/jpeg',
+      base64: 'base64EncodedString',
+      type: 'base64',
+    });
+  });
+
   describe('buildAnthropicBlock', () => {
     it('should return the content as is for text type', async () => {
       const content: UserMessageContentPart = { type: 'text', text: 'Hello!' };
@@ -52,7 +58,7 @@ describe('anthropicHelpers', () => {
         base64: null,
         type: 'url',
       });
-      vi.mocked(imageUrlToBase64).mockResolvedValue({
+      vi.mocked(imageUrlToBase64).mockResolvedValueOnce({
         base64: 'convertedBase64String',
         mimeType: 'image/jpg',
       });
@@ -82,7 +88,7 @@ describe('anthropicHelpers', () => {
         base64: null,
         type: 'url',
       });
-      vi.mocked(imageUrlToBase64).mockResolvedValue({
+      vi.mocked(imageUrlToBase64).mockResolvedValueOnce({
         base64: 'convertedBase64String',
         mimeType: 'image/png',
       });

packages/model-runtime/src/core/contextBuilders/anthropic.ts

@@ -1,8 +1,8 @@
 import Anthropic from '@anthropic-ai/sdk';
+import { imageUrlToBase64 } from '@lobechat/utils';
 import OpenAI from 'openai';
 
 import { OpenAIChatMessage, UserMessageContentPart } from '../../types';
-import { imageUrlToBase64 } from '../../utils/imageToBase64';
 import { parseDataUri } from '../../utils/uriParser';
 
 export const buildAnthropicBlock = async (

packages/model-runtime/src/core/contextBuilders/google.test.ts

@@ -1,9 +1,9 @@
 // @vitest-environment node
 import { Type as SchemaType } from '@google/genai';
+import * as imageToBase64Module from '@lobechat/utils';
 import { describe, expect, it, vi } from 'vitest';
 
 import { ChatCompletionTool, OpenAIChatMessage, UserMessageContentPart } from '../../types';
-import * as imageToBase64Module from '../../utils/imageToBase64';
 import { parseDataUri } from '../../utils/uriParser';
 import {
   buildGoogleMessage,

packages/model-runtime/src/core/contextBuilders/google.ts

@@ -5,9 +5,9 @@ import {
   Part,
   Type as SchemaType,
 } from '@google/genai';
+import { imageUrlToBase64 } from '@lobechat/utils';
 
 import { ChatCompletionTool, OpenAIChatMessage, UserMessageContentPart } from '../../types';
-import { imageUrlToBase64 } from '../../utils/imageToBase64';
 import { safeParseJSON } from '../../utils/safeParseJSON';
 import { parseDataUri } from '../../utils/uriParser';
 
@@ -64,12 +64,9 @@ export const buildGooglePart = async (
       }
 
       if (type === 'url') {
-        // For video URLs, we need to fetch and convert to base64
+        // Use imageUrlToBase64 for SSRF protection (works for any binary data including videos)
         // Note: This might need size/duration limits for practical use
-        const response = await fetch(content.video_url.url);
-        const arrayBuffer = await response.arrayBuffer();
-        const base64 = Buffer.from(arrayBuffer).toString('base64');
-        const mimeType = response.headers.get('content-type') || 'video/mp4';
+        const { base64, mimeType } = await imageUrlToBase64(content.video_url.url);
 
         return {
           inlineData: { data: base64, mimeType },

packages/model-runtime/src/core/contextBuilders/openai.test.ts

@@ -1,8 +1,8 @@
+import { imageUrlToBase64 } from '@lobechat/utils';
 import OpenAI from 'openai';
 import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
 
 import { OpenAIChatMessage } from '../../types';
-import { imageUrlToBase64 } from '../../utils/imageToBase64';
 import { parseDataUri } from '../../utils/uriParser';
 import {
   convertImageUrlToFile,
@@ -12,7 +12,9 @@ import {
 } from './openai';
 
 // 模拟依赖
-vi.mock('../../utils/imageToBase64');
+vi.mock('@lobechat/utils', () => ({
+  imageUrlToBase64: vi.fn(),
+}));
 vi.mock('../../utils/uriParser');
 
 describe('convertMessageContent', () => {

packages/model-runtime/src/core/contextBuilders/openai.ts

@@ -1,8 +1,8 @@
+import { imageUrlToBase64 } from '@lobechat/utils';
 import OpenAI, { toFile } from 'openai';
 
 import { disableStreamModels, systemToUserModels } from '../../const/models';
 import { ChatStreamPayload, OpenAIChatMessage } from '../../types';
-import { imageUrlToBase64 } from '../../utils/imageToBase64';
 import { parseDataUri } from '../../utils/uriParser';
 
 export const convertMessageContent = async (

packages/model-runtime/src/core/openaiCompatibleFactory/createImage.test.ts

@@ -1,9 +1,9 @@
 // @vitest-environment node
+import * as imageToBase64Module from '@lobechat/utils';
 import OpenAI from 'openai';
 import { beforeEach, describe, expect, it, vi } from 'vitest';
 
 import { CreateImagePayload } from '../../types/image';
-import * as imageToBase64Module from '../../utils/imageToBase64';
 import * as uriParserModule from '../../utils/uriParser';
 import { createOpenAICompatibleImage } from './createImage';
 

packages/model-runtime/src/core/openaiCompatibleFactory/createImage.ts

@@ -1,3 +1,4 @@
+import { imageUrlToBase64 } from '@lobechat/utils';
 import { cleanObject } from '@lobechat/utils/object';
 import createDebug from 'debug';
 import { RuntimeImageGenParamsValue } from 'model-bank';
@@ -5,7 +6,6 @@ import OpenAI from 'openai';
 
 import { CreateImagePayload, CreateImageResponse } from '../../types/image';
 import { getModelPricing } from '../../utils/getModelPricing';
-import { imageUrlToBase64 } from '../../utils/imageToBase64';
 import { parseDataUri } from '../../utils/uriParser';
 import { convertImageUrlToFile } from '../contextBuilders/openai';
 import { convertOpenAIImageUsage } from '../usageConverters/openai';

packages/model-runtime/src/core/openaiCompatibleFactory/index.test.ts

@@ -1,15 +1,12 @@
 // @vitest-environment node
-import {
-  AgentRuntimeErrorType,
-  ChatStreamCallbacks,
-  ChatStreamPayload,
-  LobeOpenAICompatibleRuntime,
-} from '@lobechat/model-runtime';
 import { ModelProvider } from 'model-bank';
 import OpenAI from 'openai';
 import type { Stream } from 'openai/streaming';
 import { Mock, afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
 
+import { LobeOpenAICompatibleRuntime } from '../../core/BaseAI';
+import { ChatStreamCallbacks, ChatStreamPayload } from '../../types/chat';
+import { AgentRuntimeErrorType } from '../../types/error';
 import * as debugStreamModule from '../../utils/debugStream';
 import * as openaiHelpers from '../contextBuilders/openai';
 import { createOpenAICompatibleRuntime } from './index';

packages/model-runtime/src/core/streams/openai/responsesStream.test.ts

@@ -1,6 +1,6 @@
-import { AgentRuntimeErrorType } from '@lobechat/model-runtime';
 import { describe, expect, it, vi } from 'vitest';
 
+import { AgentRuntimeErrorType } from '../../../types/error';
 import { FIRST_CHUNK_ERROR_KEY } from '../protocol';
 import { createReadableStream, readStreamChunk } from '../utils';
 import { OpenAIResponsesStream } from './responsesStream';

packages/model-runtime/src/helpers/mergeChatMethodOptions.ts

@@ -1,6 +1,7 @@
-import { ChatMethodOptions } from '@lobechat/model-runtime';
 import debug from 'debug';
 
+import { ChatMethodOptions } from '../types/chat';
+
 const log = debug('model-runtime:helpers:mergeChatMethodOptions');
 
 export const mergeMultipleChatMethodOptions = (options: ChatMethodOptions[]): ChatMethodOptions => {

packages/model-runtime/src/providers/aihubmix/index.test.ts

@@ -1,4 +1,4 @@
-// @vitest-environment edge-runtime
+// @vitest-environment node
 import { ModelProvider } from 'model-bank';
 import { beforeEach, describe, expect, it, vi } from 'vitest';
 

packages/model-runtime/src/providers/anthropic/generateObject.test.ts

@@ -1,4 +1,4 @@
-// @vitest-environment edge-runtime
+// @vitest-environment node
 import { describe, expect, it, vi } from 'vitest';
 
 import { createAnthropicGenerateObject } from './generateObject';

packages/model-runtime/src/providers/anthropic/index.test.ts

@@ -1,8 +1,8 @@
 // @vitest-environment node
-import { ChatCompletionTool, ChatStreamPayload } from '@lobechat/model-runtime';
 import { Mock, afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
 
 import * as anthropicHelpers from '../../core/contextBuilders/anthropic';
+import { ChatCompletionTool, ChatStreamPayload } from '../../types/chat';
 import * as debugStreamModule from '../../utils/debugStream';
 import { LobeAnthropicAI } from './index';
 

packages/model-runtime/src/providers/baichuan/index.test.ts

@@ -1,8 +1,8 @@
 // @vitest-environment node
-import { LobeOpenAICompatibleRuntime } from '@lobechat/model-runtime';
 import { ModelProvider } from 'model-bank';
 import { Mock, afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
 
+import { LobeOpenAICompatibleRuntime } from '../../core/BaseAI';
 import { testProvider } from '../../providerTestUtils';
 import { LobeBaichuanAI, params } from './index';
 

packages/model-runtime/src/providers/bedrock/index.test.ts

@@ -3,10 +3,10 @@ import {
   InvokeModelCommand,
   InvokeModelWithResponseStreamCommand,
 } from '@aws-sdk/client-bedrock-runtime';
-import { AgentRuntimeErrorType } from '@lobechat/model-runtime';
 import { ModelProvider } from 'model-bank';
 import { Mock, afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
 
+import { AgentRuntimeErrorType } from '../../types/error';
 import * as debugStreamModule from '../../utils/debugStream';
 import { LobeBedrockAI, experimental_buildLlama2Prompt } from './index';
 

packages/model-runtime/src/providers/bfl/createImage.test.ts

@@ -6,7 +6,7 @@ import { createBflImage } from './createImage';
 import { BflStatusResponse } from './types';
 
 // Mock external dependencies
-vi.mock('../../utils/imageToBase64', () => ({
+vi.mock('@lobechat/utils', () => ({
   imageUrlToBase64: vi.fn(),
 }));
 
@@ -187,7 +187,7 @@ describe('createBflImage', () => {
     it('should convert single imageUrl to image_prompt base64', async () => {
       // Arrange
       const { parseDataUri } = await import('../../utils/uriParser');
-      const { imageUrlToBase64 } = await import('../../utils/imageToBase64');
+      const { imageUrlToBase64 } = await import('@lobechat/utils');
       const { asyncifyPolling } = await import('../../utils/asyncifyPolling');
 
       const mockParseDataUri = vi.mocked(parseDataUri);
@@ -290,7 +290,7 @@ describe('createBflImage', () => {
     it('should convert multiple imageUrls for Kontext models', async () => {
       // Arrange
       const { parseDataUri } = await import('../../utils/uriParser');
-      const { imageUrlToBase64 } = await import('../../utils/imageToBase64');
+      const { imageUrlToBase64 } = await import('@lobechat/utils');
       const { asyncifyPolling } = await import('../../utils/asyncifyPolling');
 
       const mockParseDataUri = vi.mocked(parseDataUri);
@@ -350,7 +350,7 @@ describe('createBflImage', () => {
     it('should limit imageUrls to maximum 4 images', async () => {
       // Arrange
       const { parseDataUri } = await import('../../utils/uriParser');
-      const { imageUrlToBase64 } = await import('../../utils/imageToBase64');
+      const { imageUrlToBase64 } = await import('@lobechat/utils');
       const { asyncifyPolling } = await import('../../utils/asyncifyPolling');
 
       const mockParseDataUri = vi.mocked(parseDataUri);

packages/model-runtime/src/providers/bfl/createImage.ts

@@ -1,3 +1,4 @@
+import { imageUrlToBase64 } from '@lobechat/utils';
 import createDebug from 'debug';
 import { RuntimeImageGenParamsValue } from 'model-bank';
 
@@ -5,7 +6,6 @@ import { AgentRuntimeErrorType } from '../../types/error';
 import { CreateImagePayload, CreateImageResponse } from '../../types/image';
 import { type TaskResult, asyncifyPolling } from '../../utils/asyncifyPolling';
 import { AgentRuntimeError } from '../../utils/createError';
-import { imageUrlToBase64 } from '../../utils/imageToBase64';
 import { parseDataUri } from '../../utils/uriParser';
 import {
   BFL_ENDPOINTS,

packages/model-runtime/src/providers/cloudflare/index.test.ts

@@ -1,7 +1,7 @@
 // @vitest-environment node
-import { ChatCompletionTool } from '@lobechat/model-runtime';
 import { Mock, afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
 
+import { ChatCompletionTool } from '../../types/chat';
 import * as debugStreamModule from '../../utils/debugStream';
 import { LobeCloudflareAI } from './index';
 

packages/model-runtime/src/providers/cohere/index.test.ts

@@ -1,8 +1,8 @@
 // @vitest-environment node
-import { LobeOpenAICompatibleRuntime } from '@lobechat/model-runtime';
 import { ModelProvider } from 'model-bank';
 import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
 
+import { LobeOpenAICompatibleRuntime } from '../../core/BaseAI';
 import { testProvider } from '../../providerTestUtils';
 import { LobeCohereAI, params } from './index';
 

packages/model-runtime/src/providers/google/createImage.test.ts

@@ -1,9 +1,9 @@
-// @vitest-environment edge-runtime
+// @vitest-environment node
 import { GoogleGenAI } from '@google/genai';
+import * as imageToBase64Module from '@lobechat/utils';
 import { beforeEach, describe, expect, it, vi } from 'vitest';
 
 import { CreateImagePayload } from '../../types/image';
-import * as imageToBase64Module from '../../utils/imageToBase64';
 import { createGoogleImage } from './createImage';
 
 const provider = 'google';

packages/model-runtime/src/providers/google/createImage.ts

@@ -1,11 +1,11 @@
 import { Content, GenerateContentConfig, GoogleGenAI, Part } from '@google/genai';
+import { imageUrlToBase64 } from '@lobechat/utils';
 
 import { convertGoogleAIUsage } from '../../core/usageConverters/google-ai';
 import { CreateImagePayload, CreateImageResponse } from '../../types/image';
 import { AgentRuntimeError } from '../../utils/createError';
 import { getModelPricing } from '../../utils/getModelPricing';
 import { parseGoogleErrorMessage } from '../../utils/googleErrorParser';
-import { imageUrlToBase64 } from '../../utils/imageToBase64';
 import { parseDataUri } from '../../utils/uriParser';
 
 // Maximum number of images allowed for processing

packages/model-runtime/src/providers/google/generateObject.test.ts

@@ -1,4 +1,4 @@
-// @vitest-environment edge-runtime
+// @vitest-environment node
 import { Type as SchemaType } from '@google/genai';
 import { describe, expect, it, vi } from 'vitest';
 

packages/model-runtime/src/providers/google/index.test.ts

@@ -1,14 +1,11 @@
-// @vitest-environment edge-runtime
+// @vitest-environment node
 import { GenerateContentResponse, Tool } from '@google/genai';
-import { OpenAIChatMessage } from '@lobechat/model-runtime';
-import { ChatStreamPayload } from '@lobechat/types';
 import OpenAI from 'openai';
 import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
 
 import { LOBE_ERROR_KEY } from '../../core/streams';
 import { AgentRuntimeErrorType } from '../../types/error';
 import * as debugStreamModule from '../../utils/debugStream';
-import * as imageToBase64Module from '../../utils/imageToBase64';
 import { LobeGoogleAI, resolveModelThinkingBudget } from './index';
 
 const provider = 'google';

packages/model-runtime/src/providers/groq/index.test.ts

@@ -1,7 +1,7 @@
 // @vitest-environment node
-import { LobeOpenAICompatibleRuntime } from '@lobechat/model-runtime';
 import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
 
+import { LobeOpenAICompatibleRuntime } from '../../core/BaseAI';
 import { testProvider } from '../../providerTestUtils';
 import { AgentRuntimeErrorType } from '../../types/error';
 import { LobeGroq, params } from './index';

packages/model-runtime/src/providers/hunyuan/index.test.ts

@@ -1,8 +1,8 @@
 // @vitest-environment node
-import { LobeOpenAICompatibleRuntime } from '@lobechat/model-runtime';
 import { ModelProvider } from 'model-bank';
 import { beforeEach, describe, expect, it, vi } from 'vitest';
 
+import { LobeOpenAICompatibleRuntime } from '../../core/BaseAI';
 import { testProvider } from '../../providerTestUtils';
 import { LobeHunyuanAI, params } from './index';
 

packages/model-runtime/src/providers/minimax/createImage.test.ts

@@ -1,4 +1,4 @@
-// @vitest-environment edge-runtime
+// @vitest-environment node
 import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
 
 import { CreateImageOptions } from '../../core/openaiCompatibleFactory';

packages/model-runtime/src/providers/mistral/index.test.ts

@@ -1,7 +1,7 @@
 // @vitest-environment node
-import { LobeOpenAICompatibleRuntime } from '@lobechat/model-runtime';
 import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
 
+import { LobeOpenAICompatibleRuntime } from '../../core/BaseAI';
 import { testProvider } from '../../providerTestUtils';
 import { LobeMistralAI, params } from './index';
 

packages/model-runtime/src/providers/moonshot/index.test.ts

@@ -1,7 +1,7 @@
 // @vitest-environment node
-import { LobeOpenAICompatibleRuntime } from '@lobechat/model-runtime';
 import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
 
+import { LobeOpenAICompatibleRuntime } from '../../core/BaseAI';
 import { testProvider } from '../../providerTestUtils';
 import { LobeMoonshotAI, params } from './index';
 

packages/model-runtime/src/providers/novita/index.test.ts

@@ -1,8 +1,8 @@
 // @vitest-environment node
-import { LobeOpenAICompatibleRuntime } from '@lobechat/model-runtime';
 import { ModelProvider } from 'model-bank';
 import { Mock, afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
 
+import { LobeOpenAICompatibleRuntime } from '../../core/BaseAI';
 import { testProvider } from '../../providerTestUtils';
 import models from './fixtures/models.json';
 import { LobeNovitaAI } from './index';

packages/model-runtime/src/providers/ollama/index.test.ts

@@ -1,4 +1,5 @@
 // @vitest-environment node
+import { imageUrlToBase64 } from '@lobechat/utils';
 import { ModelProvider } from 'model-bank';
 import { Ollama } from 'ollama/browser';
 import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
@@ -9,6 +10,13 @@ import * as debugStreamModule from '../../utils/debugStream';
 import { LobeOllamaAI, params } from './index';
 
 vi.mock('ollama/browser');
+vi.mock('@lobechat/utils', async () => {
+  const actual = await vi.importActual('@lobechat/utils');
+  return {
+    ...actual,
+    imageUrlToBase64: vi.fn(),
+  };
+});
 
 // Mock the console.error to avoid polluting test output
 vi.spyOn(console, 'error').mockImplementation(() => {});
@@ -462,13 +470,13 @@ describe('LobeOllamaAI', () => {
   });
 
   describe('buildOllamaMessages', () => {
-    it('should convert OpenAIChatMessage array to OllamaMessage array', () => {
+    it('should convert OpenAIChatMessage array to OllamaMessage array', async () => {
       const messages = [
         { content: 'Hello', role: 'user' },
         { content: 'Hi there!', role: 'assistant' },
       ];
 
-      const ollamaMessages = ollamaAI['buildOllamaMessages'](messages as any);
+      const ollamaMessages = await ollamaAI['buildOllamaMessages'](messages as any);
 
       expect(ollamaMessages).toEqual([
         { content: 'Hello', role: 'user' },
@@ -476,15 +484,15 @@ describe('LobeOllamaAI', () => {
       ]);
     });
 
-    it('should handle empty message array', () => {
+    it('should handle empty message array', async () => {
       const messages: any[] = [];
 
-      const ollamaMessages = ollamaAI['buildOllamaMessages'](messages);
+      const ollamaMessages = await ollamaAI['buildOllamaMessages'](messages);
 
       expect(ollamaMessages).toEqual([]);
     });
 
-    it('should handle multiple messages with different roles', () => {
+    it('should handle multiple messages with different roles', async () => {
       const messages = [
         { content: 'Hello', role: 'system' },
         { content: 'Hi', role: 'user' },
@@ -492,7 +500,7 @@ describe('LobeOllamaAI', () => {
         { content: 'How are you?', role: 'user' },
       ];
 
-      const ollamaMessages = ollamaAI['buildOllamaMessages'](messages as any);
+      const ollamaMessages = await ollamaAI['buildOllamaMessages'](messages as any);
 
       expect(ollamaMessages).toHaveLength(4);
       expect(ollamaMessages[0].role).toBe('system');
@@ -503,26 +511,26 @@ describe('LobeOllamaAI', () => {
   });
 
   describe('convertContentToOllamaMessage', () => {
-    it('should convert string content to OllamaMessage', () => {
+    it('should convert string content to OllamaMessage', async () => {
       const message = { content: 'Hello', role: 'user' };
 
-      const ollamaMessage = ollamaAI['convertContentToOllamaMessage'](message as any);
+      const ollamaMessage = await ollamaAI['convertContentToOllamaMessage'](message as any);
 
       expect(ollamaMessage).toEqual({ content: 'Hello', role: 'user' });
     });
 
-    it('should convert text content to OllamaMessage', () => {
+    it('should convert text content to OllamaMessage', async () => {
       const message = {
         content: [{ type: 'text', text: 'Hello' }],
         role: 'user',
       };
 
-      const ollamaMessage = ollamaAI['convertContentToOllamaMessage'](message as any);
+      const ollamaMessage = await ollamaAI['convertContentToOllamaMessage'](message as any);
 
       expect(ollamaMessage).toEqual({ content: 'Hello', role: 'user' });
     });
 
-    it('should convert image_url content to OllamaMessage with images', () => {
+    it('should convert image_url content to OllamaMessage with images', async () => {
       const message = {
         content: [
           {
@@ -533,7 +541,7 @@ describe('LobeOllamaAI', () => {
         role: 'user',
       };
 
-      const ollamaMessage = ollamaAI['convertContentToOllamaMessage'](message as any);
+      const ollamaMessage = await ollamaAI['convertContentToOllamaMessage'](message as any);
 
       expect(ollamaMessage).toEqual({
         content: '',
@@ -542,7 +550,7 @@ describe('LobeOllamaAI', () => {
       });
     });
 
-    it('should ignore invalid image_url content', () => {
+    it('should ignore invalid image_url content', async () => {
       const message = {
         content: [
           {
@@ -553,7 +561,7 @@ describe('LobeOllamaAI', () => {
         role: 'user',
       };
 
-      const ollamaMessage = ollamaAI['convertContentToOllamaMessage'](message as any);
+      const ollamaMessage = await ollamaAI['convertContentToOllamaMessage'](message as any);
 
       expect(ollamaMessage).toEqual({
         content: '',
@@ -561,7 +569,7 @@ describe('LobeOllamaAI', () => {
       });
     });
 
-    it('should handle mixed text and image content', () => {
+    it('should handle mixed text and image content', async () => {
       const message = {
         content: [
           { type: 'text', text: 'First text' },
@@ -578,7 +586,7 @@ describe('LobeOllamaAI', () => {
         role: 'user',
       };
 
-      const ollamaMessage = ollamaAI['convertContentToOllamaMessage'](message as any);
+      const ollamaMessage = await ollamaAI['convertContentToOllamaMessage'](message as any);
 
       expect(ollamaMessage).toEqual({
         content: 'Second text', // Should keep latest text
@@ -587,13 +595,13 @@ describe('LobeOllamaAI', () => {
       });
     });
 
-    it('should handle content with empty text', () => {
+    it('should handle content with empty text', async () => {
       const message = {
         content: [{ type: 'text', text: '' }],
         role: 'user',
       };
 
-      const ollamaMessage = ollamaAI['convertContentToOllamaMessage'](message as any);
+      const ollamaMessage = await ollamaAI['convertContentToOllamaMessage'](message as any);
 
       expect(ollamaMessage).toEqual({
         content: '',
@@ -601,7 +609,7 @@ describe('LobeOllamaAI', () => {
       });
     });
 
-    it('should handle content with only images (no text)', () => {
+    it('should handle content with only images (no text)', async () => {
       const message = {
         content: [
           {
@@ -612,7 +620,7 @@ describe('LobeOllamaAI', () => {
         role: 'user',
       };
 
-      const ollamaMessage = ollamaAI['convertContentToOllamaMessage'](message as any);
+      const ollamaMessage = await ollamaAI['convertContentToOllamaMessage'](message as any);
 
       expect(ollamaMessage).toEqual({
         content: '',
@@ -621,7 +629,7 @@ describe('LobeOllamaAI', () => {
       });
     });
 
-    it('should handle multiple images without text', () => {
+    it('should handle multiple images without text', async () => {
       const message = {
         content: [
           {
@@ -640,7 +648,7 @@ describe('LobeOllamaAI', () => {
         role: 'user',
       };
 
-      const ollamaMessage = ollamaAI['convertContentToOllamaMessage'](message as any);
+      const ollamaMessage = await ollamaAI['convertContentToOllamaMessage'](message as any);
 
       expect(ollamaMessage).toEqual({
         content: '',
@@ -649,7 +657,10 @@ describe('LobeOllamaAI', () => {
       });
     });
 
-    it('should ignore images with invalid data URIs', () => {
+    it('should ignore images with invalid data URIs', async () => {
+      // Mock imageUrlToBase64 to simulate conversion failure for external URLs
+      vi.mocked(imageUrlToBase64).mockRejectedValue(new Error('Network error'));
+
       const message = {
         content: [
           { type: 'text', text: 'Hello' },
@@ -665,7 +676,7 @@ describe('LobeOllamaAI', () => {
         role: 'user',
       };
 
-      const ollamaMessage = ollamaAI['convertContentToOllamaMessage'](message as any);
+      const ollamaMessage = await ollamaAI['convertContentToOllamaMessage'](message as any);
 
       expect(ollamaMessage).toEqual({
         content: 'Hello',
@@ -674,7 +685,7 @@ describe('LobeOllamaAI', () => {
       });
     });
 
-    it('should handle complex interleaved content', () => {
+    it('should handle complex interleaved content', async () => {
       const message = {
         content: [
           { type: 'text', text: 'Text 1' },
@@ -692,7 +703,7 @@ describe('LobeOllamaAI', () => {
         role: 'user',
       };
 
-      const ollamaMessage = ollamaAI['convertContentToOllamaMessage'](message as any);
+      const ollamaMessage = await ollamaAI['convertContentToOllamaMessage'](message as any);
 
       expect(ollamaMessage).toEqual({
         content: 'Text 3', // Should keep latest text
@@ -701,7 +712,7 @@ describe('LobeOllamaAI', () => {
       });
     });
 
-    it('should handle assistant role with images', () => {
+    it('should handle assistant role with images', async () => {
       const message = {
         content: [
           { type: 'text', text: 'Here is the image' },
@@ -713,7 +724,7 @@ describe('LobeOllamaAI', () => {
         role: 'assistant',
       };
 
-      const ollamaMessage = ollamaAI['convertContentToOllamaMessage'](message as any);
+      const ollamaMessage = await ollamaAI['convertContentToOllamaMessage'](message as any);
 
       expect(ollamaMessage).toEqual({
         content: 'Here is the image',
@@ -722,13 +733,13 @@ describe('LobeOllamaAI', () => {
       });
     });
 
-    it('should handle system role with text', () => {
+    it('should handle system role with text', async () => {
       const message = {
         content: [{ type: 'text', text: 'You are a helpful assistant' }],
         role: 'system',
       };
 
-      const ollamaMessage = ollamaAI['convertContentToOllamaMessage'](message as any);
+      const ollamaMessage = await ollamaAI['convertContentToOllamaMessage'](message as any);
 
       expect(ollamaMessage).toEqual({
         content: 'You are a helpful assistant',
@@ -736,13 +747,13 @@ describe('LobeOllamaAI', () => {
       });
     });
 
-    it('should handle empty content array', () => {
+    it('should handle empty content array', async () => {
       const message = {
         content: [],
         role: 'user',
       };
 
-      const ollamaMessage = ollamaAI['convertContentToOllamaMessage'](message as any);
+      const ollamaMessage = await ollamaAI['convertContentToOllamaMessage'](message as any);
 
       expect(ollamaMessage).toEqual({
         content: '',

packages/model-runtime/src/providers/ollama/index.ts

@@ -1,4 +1,5 @@
 import type { ChatModelCard } from '@lobechat/types';
+import { imageUrlToBase64 } from '@lobechat/utils';
 import { ModelProvider } from 'model-bank';
 import { Ollama, Tool } from 'ollama/browser';
 import { ClientOptions } from 'openai';
@@ -61,7 +62,7 @@ export class LobeOllamaAI implements LobeRuntimeAI {
       options?.signal?.addEventListener('abort', abort);
 
       const response = await this.client.chat({
-        messages: this.buildOllamaMessages(payload.messages),
+        messages: await this.buildOllamaMessages(payload.messages),
         model: payload.model,
         options: {
           frequency_penalty: payload.frequency_penalty,
@@ -169,11 +170,13 @@ export class LobeOllamaAI implements LobeRuntimeAI {
     }
   };
 
-  private buildOllamaMessages(messages: OpenAIChatMessage[]) {
-    return messages.map((message) => this.convertContentToOllamaMessage(message));
+  private async buildOllamaMessages(messages: OpenAIChatMessage[]) {
+    return Promise.all(messages.map((message) => this.convertContentToOllamaMessage(message)));
   }
 
-  private convertContentToOllamaMessage = (message: OpenAIChatMessage): OllamaMessage => {
+  private convertContentToOllamaMessage = async (
+    message: OpenAIChatMessage,
+  ): Promise<OllamaMessage> => {
     if (typeof message.content === 'string') {
       return { content: message.content, role: message.role };
     }
@@ -183,6 +186,9 @@ export class LobeOllamaAI implements LobeRuntimeAI {
       role: message.role,
     };
 
+    // Collect image processing tasks for parallel execution
+    const imagePromises: Array<Promise<string | null> | string> = [];
+
     for (const content of message.content) {
       switch (content.type) {
         case 'text': {
@@ -191,16 +197,34 @@ export class LobeOllamaAI implements LobeRuntimeAI {
           break;
         }
         case 'image_url': {
-          const { base64 } = parseDataUri(content.image_url.url);
+          const { base64, type } = parseDataUri(content.image_url.url);
+
+          // If already base64 format, use it directly
           if (base64) {
-            ollamaMessage.images ??= [];
-            ollamaMessage.images.push(base64);
+            imagePromises.push(base64);
+          }
+          // If it's a URL, add async conversion task with error handling
+          else if (type === 'url') {
+            imagePromises.push(
+              imageUrlToBase64(content.image_url.url)
+                .then((result) => result.base64)
+                .catch(() => null), // Silently ignore failed conversions
+            );
           }
           break;
         }
       }
     }
 
+    // Process all images in parallel and filter out failed conversions
+    if (imagePromises.length > 0) {
+      const results = await Promise.all(imagePromises);
+      const validImages = results.filter((img): img is string => img !== null);
+      if (validImages.length > 0) {
+        ollamaMessage.images = validImages;
+      }
+    }
+
     return ollamaMessage;
   };
 

packages/model-runtime/src/providers/openrouter/index.test.ts

@@ -1,7 +1,7 @@
 // @vitest-environment node
-import { LobeOpenAICompatibleRuntime } from '@lobechat/model-runtime';
 import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
 
+import { LobeOpenAICompatibleRuntime } from '../../core/BaseAI';
 import { testProvider } from '../../providerTestUtils';
 import { LobeOpenRouterAI, params } from './index';
 

packages/model-runtime/src/providers/perplexity/index.test.ts

@@ -1,8 +1,8 @@
 // @vitest-environment node
-import { LobeOpenAICompatibleRuntime } from '@lobechat/model-runtime';
 import { ModelProvider } from 'model-bank';
 import { beforeEach, describe, expect, it, vi } from 'vitest';
 
+import { LobeOpenAICompatibleRuntime } from '../../core/BaseAI';
 import { testProvider } from '../../providerTestUtils';
 import { LobePerplexityAI } from './index';
 

packages/model-runtime/src/providers/ppio/index.test.ts

@@ -1,8 +1,8 @@
 // @vitest-environment node
-import { LobeOpenAICompatibleRuntime } from '@lobechat/model-runtime';
 import { ModelProvider } from 'model-bank';
 import { Mock, afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
 
+import { LobeOpenAICompatibleRuntime } from '../../core/BaseAI';
 import { testProvider } from '../../providerTestUtils';
 import models from './fixtures/models.json';
 import { LobePPIOAI } from './index';

packages/model-runtime/src/providers/qwen/createImage.test.ts

@@ -1,4 +1,4 @@
-// @vitest-environment edge-runtime
+// @vitest-environment node
 import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
 
 import { CreateImageOptions } from '../../core/openaiCompatibleFactory';

packages/model-runtime/src/providers/search1api/index.test.ts

@@ -1,7 +1,7 @@
 // @vitest-environment node
-import { LobeOpenAICompatibleRuntime } from '@lobechat/model-runtime';
 import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
 
+import { LobeOpenAICompatibleRuntime } from '../../core/BaseAI';
 import { testProvider } from '../../providerTestUtils';
 import { LobeSearch1API, params } from './index';
 

packages/model-runtime/src/providers/siliconcloud/createImage.ts

@@ -1,3 +1,4 @@
+import { imageUrlToBase64 } from '@lobechat/utils';
 import createDebug from 'debug';
 import { RuntimeImageGenParamsValue } from 'model-bank';
 
@@ -5,7 +6,6 @@ import { CreateImageOptions } from '../../core/openaiCompatibleFactory';
 import { CreateImagePayload, CreateImageResponse } from '../../types';
 import { AgentRuntimeErrorType } from '../../types/error';
 import { AgentRuntimeError } from '../../utils/createError';
-import { imageUrlToBase64 } from '../../utils/imageToBase64';
 import { parseDataUri } from '../../utils/uriParser';
 
 const log = createDebug('lobe-image:siliconcloud');

packages/model-runtime/src/providers/taichu/index.test.ts

@@ -1,9 +1,9 @@
 // @vitest-environment node
-import { LobeOpenAICompatibleRuntime } from '@lobechat/model-runtime';
 import { ModelProvider } from 'model-bank';
 import OpenAI from 'openai';
 import { Mock, afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
 
+import { LobeOpenAICompatibleRuntime } from '../../core/BaseAI';
 import { testProvider } from '../../providerTestUtils';
 import * as debugStreamModule from '../../utils/debugStream';
 import { LobeTaichuAI } from './index';

packages/model-runtime/src/providers/wenxin/index.test.ts

@@ -1,8 +1,8 @@
 // @vitest-environment node
-import { LobeOpenAICompatibleRuntime } from '@lobechat/model-runtime';
 import { ModelProvider } from 'model-bank';
 import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
 
+import { LobeOpenAICompatibleRuntime } from '../../core/BaseAI';
 import { testProvider } from '../../providerTestUtils';
 import { LobeWenxinAI, params } from './index';
 

packages/model-runtime/src/providers/zhipu/index.test.ts

@@ -1,7 +1,7 @@
 // @vitest-environment node
-import { LobeOpenAICompatibleRuntime } from '@lobechat/model-runtime';
 import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
 
+import { LobeOpenAICompatibleRuntime } from '../../core/BaseAI';
 import { testProvider } from '../../providerTestUtils';
 import { LobeZhipuAI, params } from './index';
 

packages/model-runtime/src/utils/errorResponse.test.ts

@@ -1,7 +1,7 @@
-import { AgentRuntimeErrorType } from '@lobechat/model-runtime';
 import { ChatErrorType } from '@lobechat/types';
 import { describe, expect, it, vi } from 'vitest';
 
+import { AgentRuntimeErrorType } from '../types/error';
 import { createErrorResponse } from './errorResponse';
 
 describe('createErrorResponse', () => {

packages/model-runtime/src/utils/imageToBase64.test.ts

@@ -1,91 +0,0 @@
-import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest';
-
-import { imageToBase64, imageUrlToBase64 } from './imageToBase64';
-
-describe('imageToBase64', () => {
-  let mockImage: HTMLImageElement;
-  let mockCanvas: HTMLCanvasElement;
-  let mockContext: CanvasRenderingContext2D;
-
-  beforeEach(() => {
-    mockImage = {
-      width: 200,
-      height: 100,
-    } as HTMLImageElement;
-
-    mockContext = {
-      drawImage: vi.fn(),
-    } as unknown as CanvasRenderingContext2D;
-
-    mockCanvas = {
-      width: 0,
-      height: 0,
-      getContext: vi.fn().mockReturnValue(mockContext),
-      toDataURL: vi.fn().mockReturnValue('data:image/webp;base64,mockBase64Data'),
-    } as unknown as HTMLCanvasElement;
-
-    vi.spyOn(document, 'createElement').mockReturnValue(mockCanvas);
-  });
-
-  afterEach(() => {
-    vi.restoreAllMocks();
-  });
-
-  it('should convert image to base64 with correct size and type', () => {
-    const result = imageToBase64({ img: mockImage, size: 100, type: 'image/jpeg' });
-
-    expect(document.createElement).toHaveBeenCalledWith('canvas');
-    expect(mockCanvas.width).toBe(100);
-    expect(mockCanvas.height).toBe(100);
-    expect(mockCanvas.getContext).toHaveBeenCalledWith('2d');
-    expect(mockContext.drawImage).toHaveBeenCalledWith(mockImage, 50, 0, 100, 100, 0, 0, 100, 100);
-    expect(mockCanvas.toDataURL).toHaveBeenCalledWith('image/jpeg');
-    expect(result).toBe('data:image/webp;base64,mockBase64Data');
-  });
-
-  it('should use default type when not specified', () => {
-    imageToBase64({ img: mockImage, size: 100 });
-    expect(mockCanvas.toDataURL).toHaveBeenCalledWith('image/webp');
-  });
-
-  it('should handle taller images correctly', () => {
-    mockImage.width = 100;
-    mockImage.height = 200;
-    imageToBase64({ img: mockImage, size: 100 });
-    expect(mockContext.drawImage).toHaveBeenCalledWith(mockImage, 0, 50, 100, 100, 0, 0, 100, 100);
-  });
-});
-
-describe('imageUrlToBase64', () => {
-  const mockFetch = vi.fn();
-  const mockArrayBuffer = new ArrayBuffer(8);
-
-  beforeEach(() => {
-    global.fetch = mockFetch;
-    global.btoa = vi.fn().mockReturnValue('mockBase64String');
-  });
-
-  afterEach(() => {
-    vi.restoreAllMocks();
-  });
-
-  it('should convert image URL to base64 string', async () => {
-    mockFetch.mockResolvedValue({
-      arrayBuffer: () => Promise.resolve(mockArrayBuffer),
-      blob: () => Promise.resolve(new Blob([mockArrayBuffer], { type: 'image/jpg' })),
-    });
-
-    const result = await imageUrlToBase64('https://example.com/image.jpg');
-
-    expect(mockFetch).toHaveBeenCalledWith('https://example.com/image.jpg');
-    expect(global.btoa).toHaveBeenCalled();
-    expect(result).toEqual({ base64: 'mockBase64String', mimeType: 'image/jpg' });
-  });
-
-  it('should throw an error when fetch fails', async () => {
-    const mockError = new Error('Fetch failed');
-    mockFetch.mockRejectedValue(mockError);
-
-    await expect(imageUrlToBase64('https://example.com/image.jpg')).rejects.toThrow('Fetch failed');
-  });
-});

packages/model-runtime/src/utils/imageToBase64.ts

@@ -1,62 +0,0 @@
-export const imageToBase64 = ({
-  size,
-  img,
-  type = 'image/webp',
-}: {
-  img: HTMLImageElement;
-  size: number;
-  type?: string;
-}) => {
-  const canvas = document.createElement('canvas');
-  const ctx = canvas.getContext('2d') as CanvasRenderingContext2D;
-  let startX = 0;
-  let startY = 0;
-
-  if (img.width > img.height) {
-    startX = (img.width - img.height) / 2;
-  } else {
-    startY = (img.height - img.width) / 2;
-  }
-
-  canvas.width = size;
-  canvas.height = size;
-
-  ctx.drawImage(
-    img,
-    startX,
-    startY,
-    Math.min(img.width, img.height),
-    Math.min(img.width, img.height),
-    0,
-    0,
-    size,
-    size,
-  );
-
-  return canvas.toDataURL(type);
-};
-
-export const imageUrlToBase64 = async (
-  imageUrl: string,
-): Promise<{ base64: string; mimeType: string }> => {
-  try {
-    const res = await fetch(imageUrl);
-    const blob = await res.blob();
-    const arrayBuffer = await blob.arrayBuffer();
-
-    const base64 =
-      typeof btoa === 'function'
-        ? btoa(
-            new Uint8Array(arrayBuffer).reduce(
-              (data, byte) => data + String.fromCharCode(byte),
-              '',
-            ),
-          )
-        : Buffer.from(arrayBuffer).toString('base64');
-
-    return { base64, mimeType: blob.type };
-  } catch (error) {
-    console.error('Error converting image to base64:', error);
-    throw error;
-  }
-};

packages/ssrf-safe-fetch/index.browser.ts

@@ -0,0 +1,14 @@
+/**
+ * Browser version of SSRF-safe fetch
+ * In browser environments, we simply use the native fetch API
+ * as SSRF attacks are not applicable in client-side code
+ */
+
+/**
+ * Browser-safe fetch implementation
+ * Uses native fetch API in browser environments
+ */
+// eslint-disable-next-line no-undef
+export const ssrfSafeFetch = async (url: string, options?: RequestInit): Promise<Response> => {
+  return fetch(url, options);
+};

packages/ssrf-safe-fetch/package.json

@@ -2,7 +2,14 @@
   "name": "ssrf-safe-fetch",
   "version": "1.0.0",
   "private": true,
-  "description": "",
+  "description": "SSRF-safe fetch implementation with browser/node conditional exports",
+  "exports": {
+    ".": {
+      "browser": "./index.browser.ts",
+      "node": "./index.ts",
+      "default": "./index.ts"
+    }
+  },
   "main": "index.ts",
   "scripts": {
     "test": "vitest run"

packages/utils/src/imageToBase64.ts

@@ -36,23 +36,30 @@ export const imageToBase64 = ({
   return canvas.toDataURL(type);
 };
 
+/**
+ * Convert image URL to base64
+ * Uses SSRF-safe fetch on server-side to prevent SSRF attacks
+ */
 export const imageUrlToBase64 = async (
   imageUrl: string,
 ): Promise<{ base64: string; mimeType: string }> => {
   try {
-    const res = await fetch(imageUrl);
+    const isServer = typeof window === 'undefined';
+
+    // Use SSRF-safe fetch on server-side to prevent SSRF attacks
+    const res = isServer
+      ? await import('ssrf-safe-fetch').then((m) => m.ssrfSafeFetch(imageUrl))
+      : await fetch(imageUrl);
+
     const blob = await res.blob();
     const arrayBuffer = await blob.arrayBuffer();
 
-    const base64 =
-      typeof btoa === 'function'
-        ? btoa(
-            new Uint8Array(arrayBuffer).reduce(
-              (data, byte) => data + String.fromCharCode(byte),
-              '',
-            ),
-          )
-        : Buffer.from(arrayBuffer).toString('base64');
+    // Client-side uses btoa, server-side uses Buffer
+    const base64 = isServer
+      ? Buffer.from(arrayBuffer).toString('base64')
+      : btoa(
+          new Uint8Array(arrayBuffer).reduce((data, byte) => data + String.fromCharCode(byte), ''),
+        );
 
     return { base64, mimeType: blob.type };
   } catch (error) {

packages/utils/src/index.ts

@@ -4,9 +4,9 @@ export * from './detectChinese';
 export * from './format';
 export * from './imageToBase64';
 export * from './keyboard';
+export * from './merge';
 export * from './number';
 export * from './object';
-export * from './parseModels';
 export * from './pricing';
 export * from './safeParseJSON';
 export * from './sleep';

src/features/AgentSetting/AgentTTS/SelectWithTTSPreview.tsx

@@ -1,3 +1,4 @@
+import { getMessageError } from '@lobechat/fetch-sse';
 import { ChatMessageError } from '@lobechat/types';
 import { AudioPlayer } from '@lobehub/tts/react';
 import { Alert, Button, Highlighter, Select, SelectProps } from '@lobehub/ui';
@@ -9,7 +10,6 @@ import { Flexbox } from 'react-layout-kit';
 
 import { useTTS } from '@/hooks/useTTS';
 import { TTSServer } from '@/types/agent';
-import { getMessageError } from '@/utils/fetch';
 
 interface SelectWithTTSPreviewProps extends SelectProps {
   server: TTSServer;

src/features/AgentSetting/store/action.ts

@@ -1,3 +1,4 @@
+import { MessageTextChunk } from '@lobechat/fetch-sse';
 import {
   chainPickEmoji,
   chainSummaryAgentName,
@@ -16,7 +17,6 @@ import { systemAgentSelectors } from '@/store/user/slices/settings/selectors';
 import { LobeAgentChatConfig, LobeAgentConfig } from '@/types/agent';
 import { MetaData } from '@/types/meta';
 import { SystemAgentItem } from '@/types/user/settings';
-import { MessageTextChunk } from '@/utils/fetch';
 import { merge } from '@/utils/merge';
 import { setNamespace } from '@/utils/storeDebug';
 

src/features/ChatInput/ActionBar/STT/browser.tsx

@@ -1,3 +1,4 @@
+import { getMessageError } from '@lobechat/fetch-sse';
 import { ChatMessageError } from '@lobechat/types';
 import { SpeechRecognitionOptions, useSpeechRecognition } from '@lobehub/tts/react';
 import isEqual from 'fast-deep-equal';
@@ -13,7 +14,6 @@ import { useGlobalStore } from '@/store/global';
 import { globalGeneralSelectors } from '@/store/global/selectors';
 import { useUserStore } from '@/store/user';
 import { settingsSelectors } from '@/store/user/selectors';
-import { getMessageError } from '@/utils/fetch';
 
 import CommonSTT from './common';
 

src/features/ChatInput/ActionBar/STT/openai.tsx

@@ -1,3 +1,4 @@
+import { getMessageError } from '@lobechat/fetch-sse';
 import { ChatMessageError } from '@lobechat/types';
 import { getRecordMineType } from '@lobehub/tts';
 import { OpenAISTTOptions, useOpenAISTT } from '@lobehub/tts/react';
@@ -16,7 +17,6 @@ import { useGlobalStore } from '@/store/global';
 import { globalGeneralSelectors } from '@/store/global/selectors';
 import { useUserStore } from '@/store/user';
 import { settingsSelectors } from '@/store/user/selectors';
-import { getMessageError } from '@/utils/fetch';
 
 import CommonSTT from './common';
 

src/features/Conversation/components/Extras/TTS/InitPlayer.tsx

@@ -1,3 +1,4 @@
+import { getMessageError } from '@lobechat/fetch-sse';
 import { ChatMessageError, ChatTTS } from '@lobechat/types';
 import { memo, useCallback, useEffect, useState } from 'react';
 import { useTranslation } from 'react-i18next';
@@ -5,7 +6,6 @@ import { useTranslation } from 'react-i18next';
 import { useTTS } from '@/hooks/useTTS';
 import { useChatStore } from '@/store/chat';
 import { useFileStore } from '@/store/file';
-import { getMessageError } from '@/utils/fetch';
 
 import Player from './Player';
 

src/server/globalConfig/genServerAiProviderConfig.test.ts

@@ -19,7 +19,7 @@ vi.mock('@/envs/llm', () => ({
   })),
 }));
 
-vi.mock('@/utils/parseModels', () => ({
+vi.mock('@/utils/server/parseModels', () => ({
   extractEnabledModels: vi.fn(async (providerId: string, modelString?: string) => {
     if (!modelString) return undefined;
     return [`${providerId}-model-1`, `${providerId}-model-2`];
@@ -98,7 +98,7 @@ describe('genServerAiProvidersConfig', () => {
   it('should use environment variables for model lists', async () => {
     process.env.OPENAI_MODEL_LIST = '+gpt-4,+gpt-3.5-turbo';
 
-    const { extractEnabledModels } = vi.mocked(await import('@/utils/parseModels'));
+    const { extractEnabledModels } = vi.mocked(await import('@/utils/server/parseModels'));
     extractEnabledModels.mockResolvedValue(['gpt-4', 'gpt-3.5-turbo']);
 
     const result = await genServerAiProvidersConfig({});
@@ -116,7 +116,7 @@ describe('genServerAiProvidersConfig', () => {
 
     process.env.CUSTOM_OPENAI_MODELS = '+custom-model';
 
-    const { extractEnabledModels } = vi.mocked(await import('@/utils/parseModels'));
+    const { extractEnabledModels } = vi.mocked(await import('@/utils/server/parseModels'));
 
     await genServerAiProvidersConfig(specificConfig);
 
@@ -133,7 +133,7 @@ describe('genServerAiProvidersConfig', () => {
     process.env.OPENAI_MODEL_LIST = '+gpt-4->deployment1';
 
     const { extractEnabledModels, transformToAiModelList } = vi.mocked(
-      await import('@/utils/parseModels'),
+      await import('@/utils/server/parseModels'),
     );
 
     await genServerAiProvidersConfig(specificConfig);
@@ -206,7 +206,7 @@ describe('genServerAiProvidersConfig Error Handling', () => {
       getLLMConfig: vi.fn(() => ({})),
     }));
 
-    vi.doMock('@/utils/parseModels', () => ({
+    vi.doMock('@/utils/server/parseModels', () => ({
       extractEnabledModels: vi.fn(async () => undefined),
       transformToAiModelList: vi.fn(async () => []),
     }));

src/server/globalConfig/genServerAiProviderConfig.ts

@@ -1,9 +1,9 @@
 import { ProviderConfig } from '@lobechat/types';
-import { extractEnabledModels, transformToAiModelList } from '@lobechat/utils';
 import { AiFullModelCard, ModelProvider } from 'model-bank';
 import * as AiModels from 'model-bank';
 
 import { getLLMConfig } from '@/envs/llm';
+import { extractEnabledModels, transformToAiModelList } from '@/utils/server/parseModels';
 
 interface ProviderSpecificConfig {
   enabled?: boolean;

src/services/chat/chat.test.ts

@@ -27,7 +27,7 @@ vi.stubGlobal(
 );
 
 // Mock image processing utilities
-vi.mock('@/utils/fetch', async (importOriginal) => {
+vi.mock('@lobechat/fetch-sse', async (importOriginal) => {
   const module = await importOriginal();
 
   return { ...(module as any), getMessageError: vi.fn() };
@@ -988,7 +988,7 @@ describe('ChatService', () => {
 
     beforeEach(async () => {
       // Setup common fetchSSE mock for getChatCompletion tests
-      const { fetchSSE } = await import('@/utils/fetch');
+      const { fetchSSE } = await import('@lobechat/fetch-sse');
       mockFetchSSE = vi.fn().mockResolvedValue(new Response('mock response'));
       vi.mocked(fetchSSE).mockImplementation(mockFetchSSE);
     });
@@ -1049,7 +1049,7 @@ describe('ChatService', () => {
 
     it('should return InvalidAccessCode error when enableFetchOnClient is true and auth is enabled but user is not signed in', async () => {
       // Mock fetchSSE to call onErrorHandle with the error
-      const { fetchSSE } = await import('@/utils/fetch');
+      const { fetchSSE } = await import('@lobechat/fetch-sse');
 
       const mockFetchSSEWithError = vi.fn().mockImplementation((url, options) => {
         // Simulate the error being caught and passed to onErrorHandle
@@ -1211,8 +1211,8 @@ vi.mock('../_auth', async (importOriginal) => {
 describe('ChatService private methods', () => {
   describe('getChatCompletion', () => {
     it('should merge responseAnimation styles correctly', async () => {
-      const { fetchSSE } = await import('@/utils/fetch');
-      vi.mock('@/utils/fetch', async (importOriginal) => {
+      const { fetchSSE } = await import('@lobechat/fetch-sse');
+      vi.mock('@lobechat/fetch-sse', async (importOriginal) => {
         const module = await importOriginal();
         return {
           ...(module as any),

src/services/chat/clientModelRuntime.test.ts

@@ -38,7 +38,7 @@ vi.stubGlobal(
   vi.fn(() => Promise.resolve(new Response(JSON.stringify({ some: 'data' })))),
 );
 
-vi.mock('@/utils/fetch', async (importOriginal) => {
+vi.mock('@lobechat/fetch-sse', async (importOriginal) => {
   const module = await importOriginal();
 
   return { ...(module as any), getMessageError: vi.fn() };

src/services/chat/index.ts

@@ -1,3 +1,9 @@
+import {
+  FetchSSEOptions,
+  fetchSSE,
+  getMessageError,
+  standardizeAnimationStyle,
+} from '@lobechat/fetch-sse';
 import { AgentRuntimeError, ChatCompletionErrorPayload } from '@lobechat/model-runtime';
 import { ChatErrorType, TracePayload, TraceTagMap, UIChatMessage } from '@lobechat/types';
 import { PluginRequestPayload, createHeadersWithPluginSettings } from '@lobehub/chat-plugin-sdk';
@@ -25,12 +31,6 @@ import {
 import type { ChatStreamPayload, OpenAIChatMessage } from '@/types/openai/chat';
 import { fetchWithInvokeStream } from '@/utils/electron/desktopRemoteRPCFetch';
 import { createErrorResponse } from '@/utils/errorResponse';
-import {
-  FetchSSEOptions,
-  fetchSSE,
-  getMessageError,
-  standardizeAnimationStyle,
-} from '@/utils/fetch';
 import { createTraceHeader, getTraceId } from '@/utils/trace';
 
 import { createHeaderWithAuth } from '../_auth';

src/services/chat/types.ts

@@ -1,7 +1,6 @@
+import { FetchSSEOptions } from '@lobechat/fetch-sse';
 import { TracePayload } from '@lobechat/types';
 
-import { FetchSSEOptions } from '@/utils/fetch';
-
 export interface FetchOptions extends FetchSSEOptions {
   historySummary?: string;
   signal?: AbortSignal | undefined;

src/services/models.ts

@@ -1,7 +1,8 @@
+import { getMessageError } from '@lobechat/fetch-sse';
+
 import { createHeaderWithAuth } from '@/services/_auth';
 import { aiProviderSelectors, getAiInfraStoreState } from '@/store/aiInfra';
 import { ChatModelCard } from '@/types/llm';
-import { getMessageError } from '@/utils/fetch';
 
 import { API_ENDPOINTS } from './_url';
 import { initializeWithClientStore } from './chat/clientModelRuntime';

src/utils/electron/desktopRemoteRPCFetch.ts

@@ -1,10 +1,10 @@
 import { isDesktop } from '@lobechat/const';
 import { ProxyTRPCRequestParams, dispatch, streamInvoke } from '@lobechat/electron-client-ipc';
+import { getRequestBody, headersToRecord } from '@lobechat/fetch-sse';
 import debug from 'debug';
 
 import { getElectronStoreState } from '@/store/electron';
 import { electronSyncSelectors } from '@/store/electron/selectors';
-import { getRequestBody, headersToRecord } from '@/utils/fetch';
 
 const log = debug('utils:desktopRemoteRPCFetch');
 

src/utils/server/parseModels.ts

@@ -1,9 +1,8 @@
 import { getModelPropertyWithFallback } from '@lobechat/model-runtime';
+import { merge } from '@lobechat/utils';
 import { produce } from 'immer';
 import { AiFullModelCard, AiModelType } from 'model-bank';
 
-import { merge } from './merge';
-
 /**
  * Parse model string to add or remove models.
  */

vitest.config.mts

@@ -16,6 +16,8 @@ export default defineConfig({
       // TODO: after refactor the errorResponse, we can remove it
       '@/utils/errorResponse': resolve(__dirname, './src/utils/errorResponse'),
       '@/utils/unzipFile': resolve(__dirname, './src/utils/unzipFile'),
+      '@/utils/server': resolve(__dirname, './src/utils/server'),
+      '@/utils/electron': resolve(__dirname, './src/utils/electron'),
       '@/utils': resolve(__dirname, './packages/utils/src'),
       '@/types': resolve(__dirname, './packages/types/src'),
       '@/const': resolve(__dirname, './packages/const/src'),