Connect AI agents to PostHog for autonomous product operations
Free plan available, paid tiers aboveFreemiumAI & Automation
Visit Website FreemiumVisit Website
Tracked since2026
The Bottom Line
- Price
- Free plan available, paid tiers above. Compare plans
Key facts
- Free hosted MCP server that lets AI agents perform PostHog operations using natural language prompts
- Supports many MCP-compatible clients: PostHog Code, Claude, Cursor, VS Code, Zed, Windsurf, Codex, and more
- Enables autonomous workflows like error triage, feature flag management, analytics queries, and support ticket handling
Pros
- Eliminates context switching by letting AI agents work directly in your editor
- Wide compatibility with popular MCP clients and platforms
- Generous free tier with no upfront cost to connect and use basic tools
Cons
- AI-powered tools require enabling AI data processing and may incur additional costs
- Requires an MCP-compatible client and understanding of the Model Context Protocol
What is PostHog MCP?
PostHog MCP is a free, hosted Model Context Protocol (MCP) server that enables AI agents to directly interact with your PostHog instance. Using plain text natural language prompts, agents can ship feature flags, investigate errors, run HogQL analytics queries, manage experiments, triage support tickets, configure CDP destinations, and much more — all without leaving your code editor or switching contexts. The server works with any MCP-compatible client, including PostHog Code, Claude, Cursor, VS Code, Zed, and others.
While connecting to the MCP server and calling its tools is free, some tools use large language models internally and may be billed as PostHog AI spend. The server automatically routes requests to the correct data region (US or EU) based on the authenticated account. Setup takes about 30 seconds with the PostHog Wizard, and you can choose from 20+ example prompts and multi-step recipes to get started quickly.
Preview
Key Features
- Natural language interface for PostHog operations via MCP
- Ship feature flags, run HogQL queries, and manage experiments from prompts
- Dig into stack traces, error groups, and propose fixes in-editor
- Triage support tickets and CDP destinations through AI agents
- Works with 20+ MCP-compatible clients including PostHog Code, Claude, Cursor, VS Code, and Zed
- Automatic US/EU data region routing based on authenticated account
- Free to connect; AI-powered tools billed as PostHog AI spend
- 30-second setup with the PostHog Wizard or manual configuration
Pricing Plans
Pricing checked Oct 1, 2026
| Plan | Price | Details |
|---|---|---|
| Free | Free |
+4 moreShow less
|
| Pay-as-you-go | Starts at $0/mo |
+6 moreShow less
|
Included in all plans
- Unlimited team members
- Unlimited tracked users
- API access
- SSO via Google, Github, or Gitlab
- 2FA
Is PostHog MCP worth the price?
Good value
PostHog's MCP pricing is exceptionally generous with a robust free tier covering unlimited users and team members, making it ideal for small teams and startups.
The pay-as-you-go model starts at $0 and scales with usage, offering 7-year retention and more projects, but costs can grow with heavy event volumes. This is best for product-led teams wanting to experiment with AI agents without upfront commitment.
Reviews
$99Free with your review
Write a reviewReview PostHog MCP, get a free AI guide
Share your experience and we will send you Improve Your Thinking Patterns Using ChatGPT, free.
Best PostHog MCP Alternatives
Top alternatives based on features, pricing, and user needs.
BrandfolderPaid
Manage, organize, and distribute all your brand assets
BynderPaid
Organize, manage, and distribute creative content globally
SmartlingPaid
End-to-end localization & translation services for global growth powered by AI.
EvisortPaid
AI-powered contract management software for end-to-end contract lifecycle management.
Intelligence NodePaid
AI-powered competitive intelligence for enterprise retail to optimize digital shelf performance and drive revenue.
Still deciding?
Most buyers shortlist 2 or 3 tools before committing. Pull a side-by-side comparison or browse the full alternatives shortlist below.
Explore More
PostHog MCP FAQ
How does the PostHog MCP server authenticate with my PostHog account?
The PostHog authentication server automatically routes you to the correct data region (US or EU) based on the account you log in with, so no additional region configuration is needed.
Which MCP clients are officially supported?
Supported clients include PostHog Code, Claude Desktop, Claude Code, Codex, Cursor, VS Code, Windsurf, Zed, and platforms like Lovable, Replit, and v0. The setup wizard works with all of these.
Can I use the PostHog MCP server to create and manage feature flags?
Yes, you can ship a feature flag from a prompt, manage flag assignments, and even create assignment rules for error types — all through natural language commands to your AI agent.
What types of HogQL queries can I run through the MCP server?
You can run any HogQL query, including product analytics queries, user properties, events, and information_schema tables. The agent can also surface PostHog system tables and return results in your editor.
How does the MCP server handle error tracking and triage?
The server exposes tools to retrieve top error groups, full stack traces, and metadata. Agents can propose fixes, mark issues as resolved, and assign future errors to specific teams — all from within your editor.
Is there any cost associated with using the PostHog MCP server?
Connecting to the MCP server and calling its tools is free. However, some tools use LLMs internally and may be billed as PostHog AI spend (500 free credits per month, then $0.01/credit). These AI-powered tools are only available when AI data processing is enabled in your organization settings.
Can I use the MCP server to set up CDP destinations or data pipelines?
Yes, the server includes tools for configuring CDP destinations and real-time data pipelines. You can set up a destination, manage batch exports, and trigger events — all via agent prompts.
Does the MCP server support multi-step workflows or recipes?
Yes, the documentation includes 20+ example prompts and multi-step recipes that chain together multiple tools, such as identifying errors, fixing them, and then creating a feature flag to roll out the fix.
Source: posthog.com