Agentifact assessment — independently scored, not sponsored. Last verified Mar 6, 2026.
PostHog MCP
Product analytics via MCP. Query events, feature flags, session recordings, experiments.
Solid choice for most workflows
You need your AI agents to query product analytics data, manage feature flags, and analyze user sessions without manual PostHog dashboard work or custom API integrations.
Reliable for read/write on core PostHog features with natural language; strong on analytics depth but expect API key permission limits and occasional rate limiting on high-volume queries.
You want closed-loop agentic workflows where AI can read user behavior data, decide on changes, apply feature flags or annotations, and monitor results autonomously.
Fast execution on standard ops; excels at product lifecycle tasks but write ops scoped to API key—test flag rollouts in staging first.
No OAuth Support
Relies on API keys only; no OAuth means manual key management and refresh handling in long-running agents.
API Key Permission Gotchas
Write access (flags, cohorts) requires specific key scopes—use 'MCP Server' preset; mismatched perms cause silent failures. Test read/write in MCP inspector first.
MCP standardizes agent access vs raw API's custom plumbing.
Building multi-tool AI agents needing natural language PostHog ops across Claude/Cursor/etc.
Simple scripts or non-agent apps where you control the full integration stack.
Trust Breakdown
What It Actually Does
PostHog MCP lets your AI coding agent query product analytics data—like trends, funnels, retention, and custom SQL—right from your code editor such as Cursor or VS Code. It also enables creating and managing insights and dashboards on demand.
Product analytics via MCP. Query events, feature flags, session recordings, experiments.
Fit Assessment
Best for
- ✓data-analysis
- ✓knowledge-retrieval
Score Breakdown
Protocol Support
Capabilities
Governance
- permission-scoping
- project-scoping
- audit-log