LocalSynapse
Offline AI-powered local file search MCP server for Windows. Searches inside document contents (Word, Excel, PDF, PowerPoint, HWP) using BM25 + dense vector hybrid search. 100% local, no cloud, no login, no telemetry.
README
LocalSynapse
Local file search engine with MCP server. Indexes files across all drives and provides full-text search, filename search, and semantic search — all running 100% offline.
Features
- File scanning — Indexes all fixed drives, skips cloud placeholders (OneDrive, Dropbox, etc.)
- Full-text search — BM25 ranking with Porter stemmer, Korean/English support
- Semantic search — BGE-M3 embedding model (optional, runs locally via ONNX)
- MCP server — stdio JSON-RPC server for AI assistant integration
- Desktop UI — Avalonia cross-platform GUI
Build
dotnet build LocalSynapse.v2.sln
Run
# GUI
dotnet run --project src/LocalSynapse.UI
# MCP server
dotnet run --project src/LocalSynapse.UI -- mcp
Project Structure
src/
LocalSynapse.Core/ # Models, Interfaces, DB
LocalSynapse.Pipeline/ # Scan, parse, chunk, embed
LocalSynapse.Search/ # BM25, Dense, Hybrid search
LocalSynapse.Mcp/ # MCP stdio server
LocalSynapse.UI/ # Avalonia desktop app
MCP Configuration
Add LocalSynapse as an MCP server in your Claude Desktop or Claude Code config:
{
"mcpServers": {
"localsynapse": {
"command": "C:\\Program Files\\LocalSynapse\\LocalSynapse.exe",
"args": ["mcp"],
"env": {}
}
}
}
Available MCP Tools
| Tool | Description |
|---|---|
search_files |
Search files by keyword or semantic query |
get_file_content |
Read the content of an indexed file |
list_indexed_files |
List all indexed files with filters |
get_pipeline_status |
Check indexing pipeline status |
License
Code Signing Policy
LocalSynapse releases are signed to ensure authenticity and integrity.
Free code signing provided by SignPath.io, certificate by SignPath Foundation.
Team Roles
- Authors & Reviewers: Repository Owner
- Approvers: Organization Owners
Privacy Policy
This program will not transfer any information to other networked systems unless specifically requested by the user or the person installing or operating it.
All file indexing, search, and AI embedding operations run 100% locally on your machine. No data is sent to any external server. No login or account is required.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。
mcp-server-qdrant
这个仓库展示了如何为向量搜索引擎 Qdrant 创建一个 MCP (Managed Control Plane) 服务器的示例。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。