graphrag_mcp
Provides an MCP (Model Context Protocol) server exposing tools/resources that query a hybrid GraphRAG datastore: Neo4j for graph-based context expansion and Qdrant for vector/semantic search, with hybrid retrieval combining both.
Score Breakdown
⚙ Agent Friendliness
🔒 Security
TLS is not described (README only mentions bolt for Neo4j and localhost ports). Auth appears limited to database credentials via .env; there’s no described auth/authorization for the MCP server endpoint. Secrets are presumably kept in .env (not confirmed about logging), and no explicit guidance is provided about preventing prompt/tool injection, query sanitization, or enforcing least-privilege scopes between MCP client users and the backing databases.
⚡ Reliability
Best When
You control the runtime environment (local or internal network), and want LLMs to query an existing Neo4j+Qdrant GraphRAG index via MCP tools.
Avoid When
You need multi-tenant security, public exposure without additional reverse-proxy protections, or you cannot ensure robust indexing/credential configuration.
Use Cases
- • Hybrid semantic search over document chunks with graph relationship expansion
- • Question answering where relevant context should be expanded via Neo4j relationships
- • Retrieval-augmented generation pipelines for MCP-enabled clients (e.g., Claude Desktop/Cursor)
- • Exploring and retrieving from a GraphRAG schema and Qdrant collection metadata via MCP resources
Not For
- • A hosted SaaS offering (it’s designed to run locally with your own DB instances)
- • Replacing dedicated Neo4j/Qdrant client SDKs for full administrative DB operations
- • Applications requiring strong user authentication/authorization for the MCP endpoint out of the box
Interface
Authentication
No user-facing auth model is described for the MCP server itself; authentication is limited to connecting to Neo4j/Qdrant via your configured credentials/environment variables.
Pricing
Self-hosted open-source (MIT). Costs are your infrastructure for Neo4j/Qdrant and compute (embeddings/retrieval).
Agent Metadata
Known Gotchas
- ⚠ Server depends on local Neo4j and Qdrant being reachable at the configured host/ports; misconfiguration leads to tool failures.
- ⚠ Correct operation assumes the underlying GraphRAG index has been created in both Neo4j (graph data/relationships) and Qdrant (chunk embeddings).
- ⚠ No explicit guidance is provided about rate limits, pagination strategy for large result sets, or how tool outputs should be re-requested after partial failures.
Alternatives
Full Evaluation Report
Comprehensive deep-dive: security analysis, reliability audit, agent experience review, cost modeling, competitive positioning, and improvement roadmap for graphrag_mcp.
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 2026-03-30.