AUTOMATIONSWITCH
VendorAI / ML

Qdrant MCP Server

by Qdrant

Official Qdrant vector database MCP server. Acts as a semantic memory layer on top of Qdrant: store information with metadata, retrieve via similarity search. Two tools, very small surface area, exceptionally maintained by the Qdrant team. Configurable embedding provider (fastembed default), supports remote and local Qdrant clusters.

1,373·2 tools·Released DEC 2024·Apache-2.0
uvx mcp-server-qdrant
Share:

Official Qdrant vector database MCP server. Acts as a semantic memory layer on top of Qdrant: store information with metadata, retrieve via similarity search. Two tools, very small surface area, exceptionally maintained by the Qdrant team. Pushed two days ago with 2 commits in the last 30 days; 6 releases shipped, 0.8.x line stable. The simplest path to durable agent memory backed by a real vector database. Configurable embedding provider (fastembed default with sentence-transformers/all-MiniLM-L6-v2). Supports both remote Qdrant clusters via QDRANT_URL and local databases via QDRANT_LOCAL_PATH. Read-only mode available via QDRANT_READ_ONLY for query-only deployments. FastMCP-based, distributed via PyPI as mcp-server-qdrant.

Reviewed by M. Nouriel · APR 2026

INSTALL THIS SERVER

Requires authenticationQDRANT_URL + QDRANT_API_KEY for remote Qdrant clusters, or QDRANT_LOCAL_PATH for local databases. Read-only mode via QDRANT_READ_ONLY=true.
{ "mcpServers": { "qdrant": { "command": "uvx", "args": [ "mcp-server-qdrant" ], "env": { "QDRANT_URL": "http://localhost:6333", "COLLECTION_NAME": "agent-memory", "EMBEDDING_MODEL": "sentence-transformers/all-MiniLM-L6-v2" } } } }
PrereqRequires uv installed. For remote Qdrant clusters set QDRANT_URL + QDRANT_API_KEY. For local-only operation use QDRANT_LOCAL_PATH instead. Set QDRANT_READ_ONLY=true for query-only mode. Default embedding model is sentence-transformers/all-MiniLM-L6-v2 via fastembed. Path: ~/Library/Application Support/Claude/claude_desktop_config.json (macOS).
{ "mcpServers": { "qdrant": { "command": "uvx", "args": [ "mcp-server-qdrant" ], "env": { "QDRANT_URL": "http://localhost:6333", "COLLECTION_NAME": "agent-memory", "EMBEDDING_MODEL": "sentence-transformers/all-MiniLM-L6-v2" } } } }
{ "mcpServers": { "qdrant": { "command": "uvx", "args": [ "mcp-server-qdrant" ], "env": { "QDRANT_URL": "http://localhost:6333", "COLLECTION_NAME": "agent-memory", "EMBEDDING_MODEL": "sentence-transformers/all-MiniLM-L6-v2" } } } }
{ "mcpServers": { "qdrant": { "command": "uvx", "args": [ "mcp-server-qdrant" ], "env": { "QDRANT_URL": "http://localhost:6333", "COLLECTION_NAME": "agent-memory", "EMBEDDING_MODEL": "sentence-transformers/all-MiniLM-L6-v2" } } } }
{ "mcpServers": { "qdrant": { "command": "uvx", "args": [ "mcp-server-qdrant" ], "env": { "QDRANT_URL": "http://localhost:6333", "COLLECTION_NAME": "agent-memory", "EMBEDDING_MODEL": "sentence-transformers/all-MiniLM-L6-v2" } } } }

2 TOOLS AVAILABLE

qdrant-store
Store information with optional metadata in a named Qdrant collection
Write
qdrant-find
Similarity search against a named Qdrant collection
Read

OUR ASSESSMENT

Strengths
  • Official Qdrant org publication.
  • Tiny tool surface (qdrant-store, qdrant-find) keeps the agent tool selection clean.
  • Configurable embedding model via EMBEDDING_PROVIDER and EMBEDDING_MODEL.
  • Read-only mode via QDRANT_READ_ONLY for query-only deployments.
  • Supports remote and local Qdrant via QDRANT_URL or QDRANT_LOCAL_PATH.
  • FastMCP-based with all FastMCP environment variables available.
  • Apache-2.0 license.
  • 6 releases shipped with 1,373 stars and 267 forks.
Weaknesses
  • Two tools only; teams wanting collection management or batched operations need the Qdrant client SDK.
  • Latest release v0.8.1 from December 2025; release cadence is slower than commit cadence.
  • Embedding provider currently locked to fastembed family.
Security Notes

Stores in plaintext within Qdrant (encryption at rest is a Qdrant configuration concern). For multi-tenant deployments, scope each agent to a dedicated COLLECTION_NAME. QDRANT_API_KEY pins the Qdrant cluster; rotate on credential exposure. For local-only operation use QDRANT_LOCAL_PATH so vector data stays on the host.

Best For

Teams already running Qdrant for production vector search, agents needing durable queryable memory across sessions, and read-only knowledge bases where the LLM operates in query-only mode.

TECHNICAL DETAILS

Language
python
Transport
stdiosse
Clients
Claude DesktopClaude CodeCursorVS CodeWindsurf
License
Apache-2.0
GitHub
npm
mcp-server-qdrant
Last Release
v0.8.1DEC 10, 2025
First Released
DEC 2, 2024

ADOPTION METRICS

// GitHub Stars
1,373

// Reading this1,373 stars and 267 forks confirm it as the canonical Qdrant integration path. Smithery badge in the README amplifies discovery.

// Popularity Rank
#12
Globally · #2 in AI / ML

// Reading thisSecond-ranked in ai-ml category. Cleaner Tier 1 trade-off than Pinecone for evaluation: more stars, more forks, freshly maintained.

SOURCES & VERIFICATION

We don't take any single directory's word for it. Before scoring, we cross-reference 5 public MCP sources, install the server ourselves against the clients we cover, and record when we last re-verified.

01
Discovered
Manual submission
First indexed APR 29, 2026
02
Cross-referenced
5 directories
PulseMCP, Smithery, Glama, MCP.so, Official MCP Registry
03
Verified against
Claude Desktop, Cursor, VS Code, Windsurf
Installed and tested across clients
04
Last re-checked
APR 29, 2026
Weekly re-verification
// How other directories see it

The same server, 5 different lenses. We reconcile these signals into our editorial score, which is why our number sometimes diverges from a directory-aggregate star count.

SourceTheir ratingTheir star countTheir downloadsLast synced
AutomationSwitch This page4.5editorial1,373APR 29, 2026
PulseMCP— unratedunavailableunavailableAPR 29, 2026
Smithery— unratedunavailableunavailableAPR 29, 2026
Glama— unratedunavailableunavailableAPR 29, 2026
MCP.so— unratedunavailableunavailableAPR 29, 2026
Official MCP Registry— unratedunavailableunavailableAPR 29, 2026

// Counts are directory-reported; we don't adjust them. Discrepancies usually come from different snapshot times or star-caching.

// Get in touch

DISCUSS YOUR
MCP REQUIREMENTS.

Evaluating a server, scoping an internal deployment, or working out whether MCP is the right fit at all. Start the conversation and we will point you at the right piece of the ecosystem.

Discuss Your MCP Requirements →