PostHog aims to equip every developer to build successful products.
We do this by providing the tools teams need to capture events, analyze data, record user sessions, conduct experiments, deploy new features, track errors, run surveys, provision data warehouses, observe LLM-powered features, and more, all in one platform. And we're constantly shipping new things!
PostHog is the new way to build products. Not a more modern data stack. Not faster flags. Not prettier dashboards. Everything.
A multiplayer workspace for you, your team, and your agents – with your product data as context, and PostHog tools to ship and measure. Build and edit your product | Run a fleet of agents | Turn product signals into PRs.
Tim (my cofounder and co-CEO) and I started with quite asymmetric experience.
I had previously been a VP of sales, responsible for sales, support, and account management. Tim was an insanely talented, 23-year-old engineer, and was much earlier in his career I m nearly 10 years older.
As per free tools, I recommend Posthog - Mixpanel. Both of these provide insights into your product usage and user behavior to help you make data-driven decisions and improve your products. Posthog is targeted to roles closer to web development or software related backround. Mixpanel is targeted to less technical backrounds Which tool has helped your business track user behavior and why?
PostHog Desktop (previously, PostHog Code) is an AI-powered product editor for product builders.
Problem: Most AI code editors lack real product context and start every session cold, so you’re still the one watching rollouts and catching regressions.
Solution: A multiplayer workspace where you, your team, and a fleet of agents build with your actual product data (logs, errors, session recordings, funnels, flags, experiments, tickets, etc.) as context.
What makes it different: It reads signals from production and turns them into shippable work, product signals go in, PRs come out. Unlike generic AI editors, it “knows your product” and can ship improvements while you sleep.
Key features:
Run a fleet of agents in parallel, with plan/auto modes and switchable models (Claude, Codex, open-weight).
Personal + team skills + skills marketplace, plus PostHog-maintained skills for events, flags, experiments, error tracking.
Multiplayer channels with persistent memory so agents don’t need re-briefing.
Inbox that surfaces reports and pull requests from product signals.
Benefits: Less context-switching, faster iteration from idea → plan → PR, and more outcome-oriented work (you orchestrate, agents execute).
Who it’s for: Product builders and teams who want to move from writing code / prompting outputs to orchestrating outcomes with AI agents.
Use cases: Turning in-app activity, errors, funnel drop-offs, and support signals into concrete improvements and PRs; running parallel agent tasks across a shared workspace.
Try it on macOS (or other platforms) and read the docs.
P.S. I hunt the latest and greatest launches in tech, SaaS and AI, follow to be notified →@rohanrecommends
Report
5.0
Based on 256 reviews
Review PostHog?
Reviewers largely see PostHog as a developer-friendly all-in-one platform that replaces several separate tools for analytics, session replay, feature flags, funnels, and experiments. They repeatedly praise fast setup, flexible event tracking, self-hosting, open-source transparency, and clear dashboards that help teams spot drop-off, bugs, and UX issues quickly. Founders of Roundtable, VC Boom, and Upstream describe using it to monitor adoption, funnels, and edge cases. Main drawbacks: some UI/navigation rough edges, privacy controls in replay, and a learning curve for less technical teams.
PostHog gives us a single platform for analytics, session replay, feature flags, A/B testing, and funnels — no stitching together 5 separate tools. Self-hostable, open source, and built for developers.
What needs improvement
Session replay needs privacy controls for sensitive form fields. Feature flags aren't tied to our subscription plans yet. We're not tracking match_without_full_team — that's a missed conversion signal. And our funnel
visibility from paywall view → subscription start is still blind.
vs Alternatives
Mixpanel, Amplitude, and Google Analytics. PostHog beat them on: open source, no per-event pricing, session replay built-in, and feature flags without a separate tool.
I’ve been using PostHog for about 3 months, and it has quickly become one of my favorite tools for understanding and improving our product.
What I like:
The web and product analytics dashboards are intuitive and provide deep insights into user behavior.
Session replay has been incredibly helpful in identifying UX issues and understanding how real users interact with features.
The feature flag functionality makes controlled rollouts super easy — no need for extra infrastructure or tools.
I love that everything (analytics, replays, feature flags, and experiments) is built into one platform, which saves a ton of time switching between tools.
I think they have a great platform and right abstractions of events, actions, etc. that's getting more feature rich every week and it's great that since it's still a somewhat young project they are really open to feedback and suggestions and actually are able to continually deliver improvement in the product on an ongoing basis which is pretty delightful as an end user.
Last thing that's a biggie for me too is the great culture and openness of the project and team - it's just been so easy and painless to work with them. Even if we do run into some sort of issue or feature we might want to see in the product - its as simple as a feature request in the GH repo and can have full transparency and tracking on it from there as well as see if other users feel the same way. So this lets you feel like you can easily be part of the community around PostHog and help to continually make it better.
PostHog Desktop (previously, PostHog Code) is an AI-powered product editor for product builders.
Problem: Most AI code editors lack real product context and start every session cold, so you’re still the one watching rollouts and catching regressions.
Solution: A multiplayer workspace where you, your team, and a fleet of agents build with your actual product data (logs, errors, session recordings, funnels, flags, experiments, tickets, etc.) as context.
What makes it different: It reads signals from production and turns them into shippable work, product signals go in, PRs come out. Unlike generic AI editors, it “knows your product” and can ship improvements while you sleep.
Key features:
Run a fleet of agents in parallel, with plan/auto modes and switchable models (Claude, Codex, open-weight).
Personal + team skills + skills marketplace, plus PostHog-maintained skills for events, flags, experiments, error tracking.
MCP marketplace integrations (GitHub, Slack, Linear, Figma, Stripe, Sentry, etc.).
Multiplayer channels with persistent memory so agents don’t need re-briefing.
Inbox that surfaces reports and pull requests from product signals.
Benefits: Less context-switching, faster iteration from idea → plan → PR, and more outcome-oriented work (you orchestrate, agents execute).
Who it’s for: Product builders and teams who want to move from writing code / prompting outputs to orchestrating outcomes with AI agents.
Use cases: Turning in-app activity, errors, funnel drop-offs, and support signals into concrete improvements and PRs; running parallel agent tasks across a shared workspace.
Try it on macOS (or other platforms) and read the docs.
P.S. I hunt the latest and greatest launches in tech, SaaS and AI, follow to be notified → @rohanrecommends