posthog: An open-source platform that turns errors and rage clicks into PRs

Detects signals such as errors, rage clicks, and failed queries in product data to automatically generate investigation reports and pull requests. Instead of developers manually finding and fixing issues, AI agents handle everything from diagnosis to drafting fixes, while humans only handle review and merge steps. This is called 'self-driving mode' and is the feature that makes the biggest difference compared to existing observability tools.

It is an integrated platform covering the entire product development process, not just a single observability tool. It supports product analytics, web analytics, session replay, feature flags, experiments, error tracking, logs, surveys, data warehouse, data pipelines, AI observability, and workflows in a single repository. It also includes pipeline features to sync data from external tools or send data in real time to more than 25 tools.

It can be operated directly within the editor via Slack, web, desktop app, or MCP. Connecting to MCP-compatible agents like Claude Code and Cursor allows you to instantly query PostHog data and issue tasks within your development environment. It is released under the MIT license, with code provided as open source except for the ee directory.

Suitable for development teams that need automated problem-solving based on product data and comprehensive observability. A free tier is provided up to 1 million events, 5,000 session replays, 1 million flag requests, 100,000 exceptions, and 1,500 survey responses per month, allowing adoption without cost burden in the early stages. Open-source self-hosting is recommended up to 100,000 events per month, with a recommendation to switch to the cloud service beyond that.