ragtag-mcp
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
ragtag-mcp: a local RAG setup for AI tools
Hacky MVP for local RAG search.
How to install
Clone repo
Clone this repo
cd /Users/suhdude/repos
gh repo clone weston-barger/ragtag-mcpPull down the codebases and documentation you want to index
You'll need local copies of the content you want to search. Example:
cd /Users/suhdude/repos
gh repo clone holoviz/param
gh repo clone holoviz/panel
gh repo clone bokeh/bokehSet up your rag_config.json file
Set up your rag_config.json to look like this
{
"indices": [],
"dbStoragePath": "/Users/suhdude/repos/ragtag-mcp/db",
"model": {
"embedding": "nomic-embed-text:latest",
"llm": "qwen3:latest"
}
}the dbStoragePath is where your vector database for searching will live.
Set up vitrual environment and install requirements
You'll want to set up a python venv for this repo. Run
python3 -m venv .venv
source ./.venv/bin/activate
pip3 install -r requirements.txt --upgradeInstall Ollama + accompanying models
The main.py has a command that will help you do this on OSX. Run
python main.py osx_installNOTE: you'll want to re-run the osx_install command if you change the embedding or llm models in the config.
Create your RAG search indices
Now fill our the indices section of your rag_config.json file. There is an example in example_rag_config.json. Then run
python main.py buildThis may take a a bit. See python main.py build --help to see how to build one index at a time.
Claude instructions
Install the MCP server in your Claude project
Add
{
"mcpServers": {
"RAG": {
"command": "/Users/suhdude/repos/ragtag-mcp/.venv/bin/python3",
"args": [
"/Users/suhdude/repos/ragtag-mcp/main.py",
"serve"
]
}
}
}to your .mcp.json file for your project.
Check install
Open claude an issue /mcp. You should see RAG!
Optional: Tell Claude to use the RAG search to get context
Add something to your CLAUDE.md file to tell it to use RAG to get context
(My project) relies on packages pkg1, pkg2, pkg3. Use the RAG search MCP server to understand how to use packages pkg1, pkg2, pkg3 when planning code changes. Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Related MCP Connectors
Self-hosted AI workspace: semantic RAG document search, long-term memory, and AI chat via MCP.
An MCP server for search and retrieval over a local LanceDB database
An MCP server that gives your AI access to the source code and docs of all public github repos
Related MCP Servers
- FlicenseNot gradedqualityDmaintenanceLocal MCP server that provides semantic search (RAG) over code repositories, enabling AI clients like Claude and Gemini to access project context without manual re-upload.-
- AlicenseAqualityDmaintenanceOffline RAG MCP server that indexes documents into SQLite with hybrid vector and lexical search, and enables natural language queries via Claude.41MIT
- AlicenseAqualityBmaintenanceA local RAG MCP server that indexes project documentation and code into ChromaDB for semantic search, supporting multiple tech stacks and deployment modes.201Apache 2.0
- AlicenseAqualityCmaintenanceA fully local RAG MCP server for semantic code search and code intelligence, using AST-level chunking and hybrid search to pinpoint functions, classes, and APIs. No cloud, no API keys, zero setup.16131 npm13MIT