Vector Search TF-IDF — Claude Code MCP server
Vector Search — In-Memory TF-IDF Semantic Store is an MCP server for Claude Code that indexes and searches small document collections in memory using TF-IDF and cosine similarity.
no license 3,562 installs updated 2026-04-12
claude mcp add --transport http vector-search https://vector-search--axel-belfort.run.toolsWhat Vector Search — In-Memory TF-IDF Semantic Store does
The MCP server provides an in-memory document store evaluated via TF-IDF vectorization and cosine similarity. Its primary tool, data_vector_search, takes a batch of text records alongside a search string, returning matching entries, relevance values in scores[], and the total matchCount.
This setup serves as a compact alternative to dedicated engines like Pinecone or Weaviate for local collections under 10K records and simple RAG workflows. It does not handle general web lookups; the author points to web_search_query for internet-wide retrieval.
No standard API tokens are required. Calls to https://vector-search--axel-belfort.run.tools resolve payments via x402 microtransactions on Base L2, billed at $0.005 per request.
What it helps with
- index small document batches in memory without external databases
- match queries against text snippets using TF-IDF and cosine similarity
- retrieve relevant passages for lightweight RAG pipelines under 10K documents
Example prompts for the agent
- Index these markdown notes in memory and find passages matching user authentication issues.
- Query the stored documents with
data_vector_searchto find entries relevant to payment webhooks.
How to install Vector Search — In-Memory TF-IDF Semantic Store
Platform CLI: remote server
claude mcp add --transport http vector-search https://vector-search--axel-belfort.run.toolsPlatform CLI: remote server
codex mcp add vector-search-in-memory-tf-idf-semantic-store --url https://vector-search--axel-belfort.run.toolsConfig file
{
"mcpServers": {
"vector-search-in-memory-tf-idf-semantic-store": {
"url": "https://vector-search--axel-belfort.run.tools"
}
}
}Config file
{
"servers": {
"vector-search-in-memory-tf-idf-semantic-store": {
"type": "http",
"url": "https://vector-search--axel-belfort.run.tools"
}
}
}Platform CLI: remote server
gemini mcp add -t http vector-search-in-memory-tf-idf-semantic-store https://vector-search--axel-belfort.run.toolsStep by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about Vector Search — In-Memory TF-IDF Semantic Store
How do I install Vector Search in Claude Code?
Run claude mcp add --transport http vector-search https://vector-search--axel-belfort.run.tools in your terminal to register the remote HTTP endpoint.
Does this server require an API key?
No API key is needed. Each request uses x402 micropayments on Base L2 priced at $0.005 per call.
When should you use this instead of Pinecone?
It fits small datasets under 10K documents when you need quick in-memory RAG without deploying and maintaining separate vector database infrastructure.