Fusionpact Vectordb
Give your coding agent hybrid vector plus reasoning retrieval, durable memory, and RAG without bolting together five separate services.
Overview
Fusionpact VectorDB is a Build-phase MCP server that provides hybrid vector and reasoning retrieval, agent memory, orchestration, and RAG for coding agents.
What is this MCP server?
- Hybrid vector and reasoning retrieval for RAG-style agent answers
- Agent memory and multi-agent orchestration exposed as MCP (npm fusionpact 2.1.1, stdio)
- EMBEDDING_PROVIDER switch: ollama (free), openai, or mock for local vs cloud embeds
- Single MCP server packaging vector DB concerns instead of a separate stack repo
- Published from FusionpactTech/fusionpact-vectordb on GitHub
- Server version 2.1.1
- 3 embedding provider modes: ollama, openai, mock
- 1 npm stdio package identifier: fusionpact
What problem does it solve?
Agent projects stall when context lives in chat logs instead of a searchable memory layer the MCP client can query reliably.
Who is it for?
Builders creating RAG agents, multi-step coding assistants, or internal knowledge copilots inside Claude Code or Cursor.
Skip if: Teams that only need a hosted Pinecone dashboard with no MCP integration, or pure spreadsheet analytics with no LLM retrieval.
What do I get? / Deliverables
After install, your agent can query a hybrid vector store with reasoning-backed retrieval and persist memory through one MCP stdio server.
- MCP-accessible hybrid vector and reasoning retrieval
- Agent memory persistence across sessions
- RAG-ready orchestration hooks for multi-agent flows
Recommended MCP Servers
Journey fit
Memory and retrieval belong while you are building the agent product and its MCP tool surface, before launch SEO or growth analytics. Agent-tooling is where MCP servers like Fusionpact plug into Claude Code/Cursor as the retrieval and orchestration backbone.
How it compares
Vector memory and RAG MCP server, not a governed payments or audit layer like Attestify OS.
Common Questions / FAQ
Who is Fusionpact VectorDB for?
Solo developers and indie teams building agent or RAG features who want memory and hybrid retrieval as MCP tools during implementation.
When should I use Fusionpact VectorDB?
Use it in the Build phase while wiring agent-tooling—when you need persistent memory, vector search, and orchestration before you ship and scale traffic.
How do I add Fusionpact VectorDB to my agent?
Register the npm package fusionpact (stdio transport), set EMBEDDING_PROVIDER to ollama, openai, or mock, and connect it in your MCP client config.