Vector Search API
Run in-memory vector search with TF-IDF and cosine similarity so your agent can prototype RAG-style retrieval without standing up Pinecone first.
Overview
Vector Search is a MCP server for the Build phase that performs in-memory TF-IDF vector search with cosine similarity over remote SSE with x402 micropayment.
What is this MCP server?
- Remote MCP SSE at vector-search.api.klymax402.com/mcp
- In-memory vector search using TF-IDF vectors and cosine similarity
- Suited to small corpora and RAG experiments before managed vector DBs
- x402 micropayment on hosted API v1.1.0
- Similarity: cosine on TF-IDF vectors
- Storage model: in-memory (non-durable)
- MCP v1.1.0 at vector-search.api.klymax402.com/mcp (SSE)
What problem does it solve?
You want semantic-ish document retrieval in an agent demo but managed vector databases feel heavy for a first integration.
Who is it for?
Indie agent builders validating retrieval UX on small in-memory corpora through MCP before committing to a vector DB vendor.
Skip if: Large-scale production search, multimodal embeddings, or durable indices that must survive restarts without re-ingest.
What do I get? / Deliverables
Your agent can index small text collections and return ranked similar chunks for RAG-style prompts and tooling prototypes.
- Ranked similar documents or chunks for a query
- In-memory search prototype callable from agent workflows
Recommended MCP Servers
Journey fit
Semantic retrieval wiring is an integration step while you build agent features and knowledge helpers, not a full launch or ops story on day one. Integrations captures MCP hooks that connect models to search and memory patterns during product build.
How it compares
MCP TF-IDF search API, not a hosted embedding database or full RAG orchestration skill.
Common Questions / FAQ
Who is Vector Search for?
Solo developers prototyping AI features who need quick similarity search from an MCP-connected coding agent.
When should I use Vector Search?
Use it during Build when experimenting with retrieval over notes, docs, or FAQs using TF-IDF and cosine similarity in memory.
How do I add Vector Search to my agent?
Register https://vector-search.api.klymax402.com/mcp as remote SSE in your MCP client and handle x402 payment requirements on the provider.