gnosis-mcp
Zero-config knowledge base for AI coding agents. Loads your markdown docs into a searchable database and exposes them as MCP tools — search, read, and manage documentation without leaving your editor. Works instantly with SQLite (no setup), upgrades to PostgreSQL + pgvector for hybrid semantic search. 6 MCP tools, 3 resources, FTS5 keyword search, 176 tests.
README
<!-- mcp-name: io.github.nicholasglazer/gnosis --> <div align="center">
<h1>Gnosis MCP</h1>
<p><strong>Give your AI agent a searchable knowledge base. Zero config.</strong></p>
<p> <a href="https://pypi.org/project/gnosis-mcp/"><img src="https://img.shields.io/pypi/v/gnosis-mcp?color=blue" alt="PyPI"></a> <a href="https://pypi.org/project/gnosis-mcp/"><img src="https://img.shields.io/pypi/pyversions/gnosis-mcp" alt="Python"></a> <a href="LICENSE"><img src="https://img.shields.io/badge/license-MIT-green" alt="MIT License"></a> <a href="https://github.com/nicholasglazer/gnosis-mcp/actions"><img src="https://github.com/nicholasglazer/gnosis-mcp/actions/workflows/publish.yml/badge.svg" alt="CI"></a> <a href="https://registry.modelcontextprotocol.io"><img src="https://img.shields.io/badge/MCP-Registry-blue" alt="MCP Registry"></a> </p>
<p> <a href="#quick-start">Quick Start</a> · <a href="#choose-your-backend">Backends</a> · <a href="#editor-integrations">Editor Setup</a> · <a href="#what-it-does">Tools & Resources</a> · <a href="#configuration">Configuration</a> · <a href="llms-full.txt">Full Reference</a> </p>
<a href="https://miozu.com/products/gnosis-mcp"><img src="https://miozu.com/oss/gnosis-mcp-demo.gif" alt="Gnosis MCP demo — ingest, search, serve" width="700"></a>
</div>
AI coding agents can read your source code but not your documentation. They guess at architecture, miss established patterns, and hallucinate details they could have looked up.
Gnosis MCP fixes this. Point it at a folder of docs and it creates a searchable knowledge base that any MCP-compatible AI agent can query — Claude Code, Cursor, Windsurf, Cline, and any tool that supports the Model Context Protocol.
No database server. SQLite works out of the box with keyword search, or add [embeddings] for local semantic search. Scale to PostgreSQL + pgvector when needed.
Why use this
Less hallucination. Agents search your docs before guessing. Architecture decisions, API contracts, billing rules — one tool call away instead of made up.
Lower token costs. A search returns ~600 tokens of ranked results. Reading the same docs as files costs 3,000-8,000+ tokens. On a 170-doc knowledge base (~840K tokens), that's the difference between a precise answer and a blown context window.
Docs that stay current. Add a new markdown file, run ingest, it's searchable immediately. Or use --watch to auto-re-ingest on file changes. No routing tables to maintain, no hardcoded paths to update.
Works with what you have. Gnosis MCP ingests .md, .txt, .ipynb, .toml, .csv, and .json files. Non-markdown formats are auto-converted for chunking — zero extra dependencies.
Quick Start
pip install gnosis-mcp
gnosis-mcp ingest ./docs/ # loads docs, auto-creates SQLite database
gnosis-mcp serve # starts MCP server
That's it. Your AI agent can now search your docs.
Want semantic search? Add local ONNX embeddings (no API key needed, ~23MB model):
pip install gnosis-mcp[embeddings]
gnosis-mcp ingest ./docs/ --embed # ingest + embed in one step
gnosis-mcp serve # hybrid keyword+semantic search auto-activated
Test it before connecting to an editor:
gnosis-mcp search "getting started" # keyword search
gnosis-mcp search "how does auth work" --embed # hybrid semantic+keyword
gnosis-mcp stats # see what was indexed
<details> <summary>Try without installing (uvx)</summary>
uvx gnosis-mcp ingest ./docs/
uvx gnosis-mcp serve
</details>
Editor Integrations
Gnosis MCP works with any MCP-compatible editor. Add the server config, and your AI agent gets search_docs, get_doc, and get_related tools automatically.
Claude Code
Add to .claude/mcp.json:
{
"mcpServers": {
"docs": {
"command": "gnosis-mcp",
"args": ["serve"]
}
}
}
Or install as a Claude Code plugin for a richer experience with slash commands.
Cursor
Add to .cursor/mcp.json:
{
"mcpServers": {
"docs": {
"command": "gnosis-mcp",
"args": ["serve"]
}
}
}
Windsurf
Add to ~/.codeium/windsurf/mcp_config.json:
{
"mcpServers": {
"docs": {
"command": "gnosis-mcp",
"args": ["serve"]
}
}
}
VS Code (GitHub Copilot)
Add to .vscode/mcp.json in your workspace:
{
"servers": {
"docs": {
"command": "gnosis-mcp",
"args": ["serve"]
}
}
}
Also discoverable via the VS Code MCP gallery — search @mcp gnosis in the Extensions view.
Enterprise: Your org admin needs the "MCP servers in Copilot" policy enabled. Free/Pro/Pro+ plans work without this.
JetBrains (IntelliJ, PyCharm, WebStorm)
Go to Settings > Tools > AI Assistant > MCP Servers, click +, and add:
- Name:
docs - Command:
gnosis-mcp - Arguments:
serve
Cline
Open Cline MCP settings panel and add the same server config.
Other MCP clients
Any tool that supports the Model Context Protocol works — including Zed, Neovim (via plugins), and custom agents. The server communicates over stdio by default, or Streamable HTTP for remote deployment:
gnosis-mcp serve --transport streamable-http --host 0.0.0.0 --port 8000
# Remote clients connect to http://your-server:8000/mcp
Choose Your Backend
| SQLite (default) | SQLite + embeddings | PostgreSQL | |
|---|---|---|---|
| Install | pip install gnosis-mcp |
pip install gnosis-mcp[embeddings] |
pip install gnosis-mcp[postgres] |
| Config | Nothing | Nothing | Set DATABASE_URL |
| Search | FTS5 keyword (BM25) | Hybrid keyword + semantic (RRF) | tsvector + pgvector hybrid |
| Embeddings | None | Local ONNX (23MB, no API key) | Any provider + HNSW index |
| Multi-table | No | No | Yes (UNION ALL) |
| Best for | Quick start, keyword-only | Semantic search without a server | Production, large doc sets |
Auto-detection: Set DATABASE_URL to postgresql://... and it uses PostgreSQL. Don't set it and it uses SQLite. Override with GNOSIS_MCP_BACKEND=sqlite|postgres.
<details> <summary>PostgreSQL setup</summary>
pip install gnosis-mcp[postgres]
export GNOSIS_MCP_DATABASE_URL="postgresql://user:pass@localhost:5432/mydb"
gnosis-mcp init-db # create tables + indexes
gnosis-mcp ingest ./docs/ # load your markdown
gnosis-mcp serve
For hybrid semantic+keyword search, also enable pgvector:
CREATE EXTENSION IF NOT EXISTS vector;
Then backfill embeddings:
gnosis-mcp embed # via OpenAI (default)
gnosis-mcp embed --provider ollama # or use local Ollama
</details>
Claude Code Plugin
For Claude Code users, install as a plugin to get the MCP server plus slash commands:
claude plugin marketplace add nicholasglazer/gnosis-mcp
claude plugin install gnosis
This gives you:
| Component | What you get |
|---|---|
| MCP server | gnosis-mcp serve — auto-configured |
/gnosis:search |
Search docs with keyword or --semantic hybrid mode |
/gnosis:status |
Health check — connectivity, doc stats, troubleshooting |
/gnosis:manage |
CRUD — add, delete, update metadata, bulk embed |
The plugin works with both SQLite and PostgreSQL backends.
<details> <summary>Manual setup (without plugin)</summary>
Add to .claude/mcp.json:
{
"mcpServers": {
"gnosis": {
"command": "gnosis-mcp",
"args": ["serve"]
}
}
}
For PostgreSQL, add "env": {"GNOSIS_MCP_DATABASE_URL": "postgresql://..."}.
</details>
What It Does
Gnosis MCP exposes 6 tools and 3 resources over MCP. Your AI agent calls these automatically when it needs information from your docs.
Tools
| Tool | What it does | Mode |
|---|---|---|
search_docs |
Search by keyword or hybrid semantic+keyword | Read |
get_doc |
Retrieve a full document by path | Read |
get_related |
Find linked/related documents | Read |
upsert_doc |
Create or replace a document | Write |
delete_doc |
Remove a document and its chunks | Write |
update_metadata |
Change title, category, tags | Write |
Read tools are always available. Write tools require GNOSIS_MCP_WRITABLE=true.
Resources
| URI | Returns |
|---|---|
gnosis://docs |
All documents — path, title, category, chunk count |
gnosis://docs/{path} |
Full document content |
gnosis://categories |
Categories with document counts |
How search works
# Keyword search — works on both SQLite and PostgreSQL
gnosis-mcp search "stripe webhook"
# Hybrid search — keyword + semantic similarity (PostgreSQL + embeddings)
gnosis-mcp search "how does billing work" --embed
# Filtered — narrow results to a specific category
gnosis-mcp search "auth" -c guides
When called via MCP, the agent passes a query string for keyword search. On PostgreSQL with embeddings, it can also pass query_embedding for hybrid mode that combines keyword matching with semantic similarity.
Search results include a highlight field with matched terms wrapped in <mark> tags for context-aware snippets (FTS5 snippet() on SQLite, ts_headline() on PostgreSQL).
Embeddings
Embeddings enable semantic search — finding docs by meaning, not just keywords.
1. Local ONNX (recommended for SQLite) — zero-config, no API key needed:
pip install gnosis-mcp[embeddings]
gnosis-mcp ingest ./docs/ --embed # ingest + embed in one step
gnosis-mcp embed # or embed existing chunks separately
Uses MongoDB/mdbr-leaf-ir (~23MB quantized, Apache 2.0). Auto-downloads on first run. Customize with GNOSIS_MCP_EMBED_MODEL.
2. Remote providers — OpenAI, Ollama, or any OpenAI-compatible endpoint:
gnosis-mcp embed --provider openai # requires GNOSIS_MCP_EMBED_API_KEY
gnosis-mcp embed --provider ollama # uses local Ollama server
3. Pre-computed vectors — pass embeddings to upsert_doc or query_embedding to search_docs from your own pipeline.
Hybrid search — when embeddings are available, search automatically combines keyword (BM25) and semantic (cosine) results using Reciprocal Rank Fusion (RRF). Works on both SQLite (via sqlite-vec) and PostgreSQL (via pgvector).
Configuration
All settings via environment variables. Nothing required for SQLite — it works with zero config.
| Variable | Default | Description |
|---|---|---|
GNOSIS_MCP_DATABASE_URL |
SQLite auto | PostgreSQL URL or SQLite file path |
GNOSIS_MCP_BACKEND |
auto |
Force sqlite or postgres |
GNOSIS_MCP_WRITABLE |
false |
Enable write tools (upsert_doc, delete_doc, update_metadata) |
GNOSIS_MCP_TRANSPORT |
stdio |
Server transport: stdio or sse |
GNOSIS_MCP_SCHEMA |
public |
Database schema (PostgreSQL only) |
GNOSIS_MCP_CHUNKS_TABLE |
documentation_chunks |
Table name for chunks |
GNOSIS_MCP_SEARCH_FUNCTION |
— | Custom search function (PostgreSQL only) |
GNOSIS_MCP_EMBEDDING_DIM |
1536 |
Vector dimension for init-db |
<details> <summary>All variables</summary>
Search & chunking: GNOSIS_MCP_CONTENT_PREVIEW_CHARS (200), GNOSIS_MCP_CHUNK_SIZE (4000), GNOSIS_MCP_SEARCH_LIMIT_MAX (20).
Connection pool (PostgreSQL): GNOSIS_MCP_POOL_MIN (1), GNOSIS_MCP_POOL_MAX (3).
Webhooks: GNOSIS_MCP_WEBHOOK_URL, GNOSIS_MCP_WEBHOOK_TIMEOUT (5s). Set a URL to receive POST notifications when documents are created, updated, or deleted.
Embeddings: GNOSIS_MCP_EMBED_PROVIDER (openai/ollama/custom/local), GNOSIS_MCP_EMBED_MODEL (text-embedding-3-small for remote, MongoDB/mdbr-leaf-ir for local), GNOSIS_MCP_EMBED_DIM (384, Matryoshka truncation dimension for local provider), GNOSIS_MCP_EMBED_API_KEY, GNOSIS_MCP_EMBED_URL (custom endpoint), GNOSIS_MCP_EMBED_BATCH_SIZE (50).
Column overrides (for connecting to existing tables with non-standard column names): GNOSIS_MCP_COL_FILE_PATH, GNOSIS_MCP_COL_TITLE, GNOSIS_MCP_COL_CONTENT, GNOSIS_MCP_COL_CHUNK_INDEX, GNOSIS_MCP_COL_CATEGORY, GNOSIS_MCP_COL_AUDIENCE, GNOSIS_MCP_COL_TAGS, GNOSIS_MCP_COL_EMBEDDING, GNOSIS_MCP_COL_TSV, GNOSIS_MCP_COL_SOURCE_PATH, GNOSIS_MCP_COL_TARGET_PATH, GNOSIS_MCP_COL_RELATION_TYPE.
Links table: GNOSIS_MCP_LINKS_TABLE (documentation_links).
Logging: GNOSIS_MCP_LOG_LEVEL (INFO).
</details>
<details> <summary>Custom search function (PostgreSQL)</summary>
Delegate search to your own PostgreSQL function for custom ranking:
CREATE FUNCTION my_schema.my_search(
p_query_text text,
p_categories text[],
p_limit integer
) RETURNS TABLE (
file_path text, title text, content text,
category text, combined_score double precision
) ...
GNOSIS_MCP_SEARCH_FUNCTION=my_schema.my_search
</details>
<details> <summary>Multi-table mode (PostgreSQL)</summary>
Query across multiple doc tables:
GNOSIS_MCP_CHUNKS_TABLE=documentation_chunks,api_docs,tutorial_chunks
All tables must share the same schema. Reads use UNION ALL. Writes target the first table.
</details>
CLI Reference
gnosis-mcp ingest <path> [--dry-run] [--force] [--embed] Load files (--force to re-ingest unchanged)
gnosis-mcp serve [--transport stdio|sse] [--ingest PATH] [--watch PATH] Start MCP server (--watch for live reload)
gnosis-mcp search <query> [-n LIMIT] [-c CAT] [--embed] Search (--embed for hybrid semantic+keyword)
gnosis-mcp stats Show document, chunk, and embedding counts
gnosis-mcp check Verify database connection + sqlite-vec status
gnosis-mcp embed [--provider P] [--model M] [--dry-run] Backfill embeddings (auto-detects local provider)
gnosis-mcp init-db [--dry-run] Create tables + indexes manually
gnosis-mcp export [-f json|markdown|csv] [-c CAT] Export documents
gnosis-mcp diff <path> Show what would change on re-ingest
How ingestion works
gnosis-mcp ingest scans a directory for supported files (.md, .txt, .ipynb, .toml, .csv, .json) and loads them into the database:
- Multi-format — Markdown native;
.txt,.ipynb,.toml,.csv,.jsonauto-converted (stdlib only). Optional:.rst(pip install gnosis-mcp[rst]),.pdf(pip install gnosis-mcp[pdf]) - Smart chunking — splits by H2 headings (H3/H4 for oversized sections), never splits inside fenced code blocks or tables
- Frontmatter support — extracts
title,category,audience,tagsfrom YAML frontmatter - Auto-linking —
relates_toin frontmatter creates bidirectional links (queryable viaget_related) - Auto-categorization — infers category from the parent directory name
- Incremental updates — content hashing skips unchanged files on re-run (
--forceto override) - Watch mode —
gnosis-mcp serve --watch ./docs/auto-re-ingests on file changes - Dry run — preview what would be indexed with
--dry-run
Available on
Gnosis MCP is listed on the Official MCP Registry (which feeds the VS Code MCP gallery and GitHub Copilot), PyPI, and major MCP directories including mcp.so, Glama, and cursor.directory.
Architecture
src/gnosis_mcp/
├── backend.py DocBackend protocol + create_backend() factory
├── pg_backend.py PostgreSQL — asyncpg, tsvector, pgvector
├── sqlite_backend.py SQLite — aiosqlite, FTS5, sqlite-vec hybrid search (RRF)
├── sqlite_schema.py SQLite DDL — tables, FTS5, triggers, vec0 virtual table
├── config.py Config from env vars, backend auto-detection
├── db.py Backend lifecycle + FastMCP lifespan
├── server.py FastMCP server — 6 tools, 3 resources, auto-embed queries
├── ingest.py File scanner + converters — multi-format, smart chunking (H2/H3/H4)
├── watch.py File watcher — mtime polling, auto-re-ingest on changes
├── schema.py PostgreSQL DDL — tables, indexes, search functions
├── embed.py Embedding providers — OpenAI, Ollama, custom, local ONNX
├── local_embed.py Local ONNX embedding engine — HuggingFace model download
└── cli.py CLI — serve, ingest, search, embed, stats, check
AI-Friendly Docs
These files are optimized for AI agents to consume:
| File | Purpose |
|---|---|
llms.txt |
Quick overview — what it does, tools, config |
llms-full.txt |
Complete reference in one file |
llms-install.md |
Step-by-step installation guide |
Development
git clone https://github.com/nicholasglazer/gnosis-mcp.git
cd gnosis-mcp
python -m venv .venv && source .venv/bin/activate
pip install -e ".[dev]"
pytest # 240+ tests, no database needed
ruff check src/ tests/
All tests run without a database. Keep it that way.
Good first contributions: new embedding providers, export formats, ingestion for RST/HTML/PDF (via optional extras). Open an issue first for larger changes.
Sponsors
If Gnosis MCP saves you time, consider sponsoring the project.
License
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。
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。
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。
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。
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。
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。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
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。