Qdrant MCP Server vs SQLite MCP Server

A side-by-side look at pricing, capabilities, pros, cons, and our editorial scores.

	Qdrant MCP Server MCP Servers	SQLite MCP Server MCP Servers
Tagline	Official Qdrant MCP server that turns a vector database into a semantic memory layer for Claude, Cursor, Windsurf, and any MCP client.	Reference MCP server for querying and analyzing SQLite databases through Claude and other MCP clients.
Category	MCP Servers	MCP Servers
Pricing	Free· Free and open source (Apache-2.0). Qdrant itself can be self-hosted for free or used via Qdrant Cloud (free tier available, paid plans from ~$25/mo for managed clusters).	Free· Free / open source (MIT). No usage fees; runs locally against a SQLite file you control.
Model	FastEmbed (default: sentence-transformers/all-MiniLM-L6-v2); pairs with any MCP-capable LLM such as Claude 3.5/4, GPT-4o, or local models	—
Editorial score	—	—
Use cases	Persistent memory for Claude Desktop agentsSemantic code snippet search in Cursor and WindsurfPrivate documentation retrieval for internal LLM copilotsTeam knowledge base backed by Qdrant CloudLocal offline vector memory via QDRANT_LOCAL_PATHRead-only knowledge lookup skill for customer-support agentsCross-session context store for autonomous coding agents	Ad-hoc SQL exploration of a local SQLite database from Claude DesktopPrototyping agentic business-intelligence workflowsTeaching an LLM to write SQL against an introspected schemaBuilding a running insights memo across a multi-turn analysis sessionReference implementation for authoring a custom MCP serverWiring a local SQLite cache into a larger MCP-based agent stackQuick schema documentation via `list_tables` and `describe_table`Lightweight data prep and table creation inside a chat session
Pros	Official, actively maintained by the Qdrant team with 1.4k+ stars and frequent releases Two-tool surface (store/find) is small enough that models actually use it correctly Bundled FastEmbed means no separate OpenAI/Cohere embedding key is required to get started Configurable tool descriptions let you rebrand the same server as memory, code search, or docs lookup Works across the major MCP clients: Claude Desktop, Cursor, Windsurf, VS Code, and custom agents Apache-2.0 with local, Docker, and uvx install paths including a fully offline QDRANT_LOCAL_PATH mode Read-only mode makes it safe to expose a curated knowledge base without letting the model write to it	Zero-config local database access for any MCP client — point it at a .db file and Claude can query, schema-introspect and write immediately. Clean, minimal tool surface (six tools) that maps cleanly to how an LLM actually reasons about a database. Novel `append_insight` + `memo://insights` pattern gives the model a persistent scratchpad for multi-turn analysis. MIT-licensed and open source, so it doubles as a canonical example for building your own MCP server. Multiple install paths — uv, Docker, VS Code one-click — cover most developer setups. No API keys, no cloud, no per-query cost; everything runs on your machine against a file you own.
Cons	Only FastEmbed is supported today, so you cannot plug in OpenAI, Cohere, or Voyage embeddings without forking Just two tools: no filtering, hybrid search, payload updates, or collection management surfaced to the model Single active collection per server process; multi-collection agents need multiple server instances or wrapping Assumes you already run and secure a Qdrant instance (self-hosted or Cloud) — not a turnkey managed product Chunking, ingestion pipelines, and re-ranking are entirely your problem; this is a thin bridge, not a RAG framework	Repository was archived on 29 May 2025 — no more upstream fixes, security patches or new features. Exposes `write_query` and `create_table` to the model, so a careless prompt can mutate or drop data; there is no built-in read-only mode or row-level safety. SQLite-only — no Postgres, MySQL, DuckDB or cloud-warehouse support; you need a different MCP server for those. No authentication, quota or audit layer; intended for local single-user use, not shared/multi-tenant deployments. Insight-memo state lives in the running server process, so it does not survive restarts or multiple concurrent clients cleanly.
Website	github.com	github.com

Pick Qdrant MCP Server if

✅ Official, actively maintained by the Qdrant team with 1.4k+ stars and frequent releases
✅ Two-tool surface (store/find) is small enough that models actually use it correctly
✅ Bundled FastEmbed means no separate OpenAI/Cohere embedding key is required to get started
✅ Configurable tool descriptions let you rebrand the same server as memory, code search, or docs lookup

Pick SQLite MCP Server if

✅ Zero-config local database access for any MCP client — point it at a .db file and Claude can query, schema-introspect and write immediately.
✅ Clean, minimal tool surface (six tools) that maps cleanly to how an LLM actually reasons about a database.
✅ Novel `append_insight` + `memo://insights` pattern gives the model a persistent scratchpad for multi-turn analysis.
✅ MIT-licensed and open source, so it doubles as a canonical example for building your own MCP server.

Compare a different pair →