mcp-graphiti

Provides an MCP (Model Context Protocol) server setup for Graphiti-based knowledge graph extraction and storage in Neo4j. It adds a CLI that generates a root MCP server plus multiple project-specific MCP servers (via Docker Compose) so multiple knowledge graphs can share the same Neo4j database while staying isolated by project configuration.

Evaluated Mar 30, 2026 (21d ago)
Repo ↗ Ai Ml mcp graphiti neo4j knowledge-graph docker-compose python self-hosted sse cli
⚙ Agent Friendliness
40
/ 100
Can an agent use this?
🔒 Security
33
/ 100
Is it safe for agents?
⚡ Reliability
32
/ 100
Does it work consistently?

Score Breakdown

⚙ Agent Friendliness

MCP Quality
80
Documentation
55
Error Messages
0
Auth Simplicity
20
Rate Limits
10

🔒 Security

TLS Enforcement
40
Auth Strength
20
Scope Granularity
10
Dep. Hygiene
45
Secret Handling
55

README includes a warning about using the default Neo4j password and an explicit danger switch to destroy the entire graph. However, there is no documented authentication for MCP access, and the endpoints appear to be localhost-oriented; security posture likely depends on local network access controls, Docker configuration, and protecting .env values.

⚡ Reliability

Uptime/SLA
0
Version Stability
40
Breaking Changes
35
Error Recovery
55
AF Security Reliability

Best When

You want self-hosted MCP endpoints for knowledge-graph operations with multiple project configurations sharing one Neo4j instance.

Avoid When

You cannot provide secure Neo4j credentials and environment variables, or you need clear, documented request authentication and rate-limit behavior for external clients.

Use Cases

  • Create and manage multiple knowledge-graph projects with isolated extraction/entity definitions
  • Run MCP-compatible tools against Graphiti/Neo4j via SSE endpoints
  • Use Docker-based multi-project orchestration for local or self-hosted deployments
  • Reload/iterate on project MCP server configuration and prompts

Not For

  • Production deployments that require strong, documented authentication/authorization controls out of the box
  • Use cases that need a managed/hosted service SLA
  • Teams that require documented REST/GraphQL/OpenAPI contracts beyond MCP

Interface

REST API
No
GraphQL
No
gRPC
No
MCP Server
Yes
SDK
No
Webhooks
No

Authentication

Methods: Environment-variable based configuration for Neo4j connection (credentials in .env) No explicit mention of HTTP request auth for the MCP endpoints
OAuth: No Scopes: No

README does not describe any authentication/authorization for accessing the MCP SSE endpoints. Security effectively relies on network controls and protecting environment variables/Neo4j credentials.

Pricing

Free tier: No
Requires CC: No

Costs are primarily infrastructure plus any OpenAI/LLM usage configured via environment variables (not priced/quantified in README).

Agent Metadata

Pagination
none
Idempotent
False
Retry Guidance
Not documented

Known Gotchas

  • Root server is on port 8000; project containers start at 8001—agents should discover the correct SSE endpoint via the generated .cursor/mcp.json when possible
  • Default Neo4j password (password) is refused unless GRAPHITI_ENV=dev—automation should ensure proper env configuration
  • Disabling/using NEO4J_DESTROY_ENTIRE_GRAPH=true can irreversibly wipe graphs; avoid enabling this in automated environments
  • The CLI-generated orchestration suggests a container lifecycle; agents should account for startup delays and ensure services are up before issuing MCP calls

Alternatives

Full Evaluation Report

Comprehensive deep-dive: security analysis, reliability audit, agent experience review, cost modeling, competitive positioning, and improvement roadmap for mcp-graphiti.

AI-powered analysis · PDF + markdown · Delivered within 30 minutes

$99

Package Brief

Quick verdict, integration guide, cost projections, gotchas with workarounds, and alternatives comparison.

Delivered within 10 minutes

$3

Score Monitoring

Get alerted when this package's AF, security, or reliability scores change significantly. Stay ahead of regressions.

Continuous monitoring

$3/mo

Scores are editorial opinions as of 2026-03-30.

8642
Packages Evaluated
17761
Need Evaluation
586
Need Re-evaluation
Community Powered