mcp-memory-server
An MCP (Model Context Protocol) server that provides a memory storage service with tools to store, retrieve (semantic similarity), search by tags, and delete memories. It communicates via stdio (MCP over JSON-RPC transport). Storage backends include in-memory and ChromaDB; embedding models include a dummy generator and an ONNX Runtime-based embedding generator.
Score Breakdown
⚙ Agent Friendliness
🔒 Security
Transport is stdio (no TLS applies). The README does not mention authentication/authorization or any access control. Configuration is via environment variables; no guidance is provided about secret management (though no secrets are explicitly required). Dependency hygiene and vulnerability posture cannot be assessed from the provided data.
⚡ Reliability
Best When
Used as a local/controlled MCP stdio server where the client (e.g., an MCP-compatible desktop/editor integration) orchestrates tool calls for memory operations.
Avoid When
Avoid for internet-exposed or multi-tenant services unless you wrap it with a trusted gateway, because the README describes no authentication/authorization and shows tools that can store/retrieve/delete data.
Use Cases
- • Agent memory for conversational assistants (store and retrieve relevant past items)
- • Tag-based knowledge browsing (search memories by metadata/tags)
- • Semantic retrieval for RAG-style workflows using an MCP client
- • Local/dev prototyping of memory tooling without external auth
Not For
- • Multi-tenant production deployments needing strong authentication/authorization
- • Use cases requiring a REST/HTTP API or webhooks
- • Environments where running ONNX Runtime models is not acceptable
- • Systems that require documented pagination, rate limits, and robust API error contracts beyond MCP/tool calls
Interface
Authentication
No authentication/authorization is described. The server runs as an MCP stdio process; any access control would need to be enforced by the MCP client/gateway or at the process/network boundary.
Pricing
No pricing model is provided in the supplied data.
Agent Metadata
Known Gotchas
- ⚠ No documented auth: an agent/operator should assume the process has full access to the configured storage backend.
- ⚠ Embedding/Chroma configuration is environment-variable driven; incorrect model/storage settings may cause tool failures.
- ⚠ Pagination/limits are not described for retrieval/search results (agent may need to manage result sizes/behavior based on tool outputs).
Alternatives
Full Evaluation Report
Comprehensive deep-dive: security analysis, reliability audit, agent experience review, cost modeling, competitive positioning, and improvement roadmap for mcp-memory-server.
AI-powered analysis · PDF + markdown · Delivered within 30 minutes
Package Brief
Quick verdict, integration guide, cost projections, gotchas with workarounds, and alternatives comparison.
Delivered within 10 minutes
Score Monitoring
Get alerted when this package's AF, security, or reliability scores change significantly. Stay ahead of regressions.
Continuous monitoring
Scores are editorial opinions as of 2026-04-04.