mcp-server-qdrant
Qdrant-backed semantic memory MCP for storing and retrieving memories via vector search, with a read-only mode.
🛡️ AgentReady threat assessment
MAESTRO 7-layer threat model + OWASP AIVSS risk score for mcp-server-qdrant, derived from its capabilities.
AIVSS 7.0 · High
View MAESTRO 7-layer threat model →Overview
mcp-server-qdrant acts as a semantic memory layer, storing and retrieving memories through Qdrant vector search with configurable embedding providers and an optional read-only flag. Security surface: it holds QDRANT_URL/API_KEY and an embedding provider key, and the stored memories themselves are a prompt-injection vector when read back into an agent's context.
Key features
- Store and retrieve vector memories in Qdrant
- Pluggable embedding providers
- Optional read-only mode
- Configurable collection and search limit
Use cases
- Give an agent persistent semantic memory
- Back a RAG pipeline with Qdrant