← back to blog

i thought you were building posthog, why twig?

January 2026

i thought you were building posthog, why twig?

Fair question. PostHog is doing great - over 100,000 teams use us, we're profitable, and we're still building features every week. So why start something new?

we saw something coming

For years, PostHog has been building the best tools for understanding user behavior. Analytics, session replay, feature flags, experiments, surveys - all the data you need to build better products.

But here's what we noticed: the workflow is broken. You look at a funnel, spot a drop-off, open your code editor, make a change, deploy, wait for data, repeat. Humans are the glue between insight and action.

That glue is about to dissolve.

agents change everything

AI agents are getting scary good at writing code. Not perfect, but good enough to handle the routine stuff - fixing bugs, improving copy, tweaking layouts, running experiments. The kind of work that takes up 80% of a product team's time.

We realized that PostHog's data - all those insights about user behavior - is exactly what these agents need to make good decisions. An agent with access to session recordings can spot UX problems. An agent with funnel data can prioritize what to fix. An agent with experiment results can iterate automatically.

PostHog becomes infinitely more valuable when agents can use it directly.

twig is posthog's future

Think of Twig as the action layer for PostHog. PostHog tells you what's happening. Twig does something about it. They're complementary - Twig needs PostHog's data to be useful, and PostHog becomes more powerful when Twig can act on its insights.

We're not abandoning PostHog. We're extending it into the world of autonomous development. The same team, the same mission - helping you build better products. Just with a lot more automation.

why not just add this to posthog?

We thought about it. But code editors are a different product category with different users and different expectations. Trying to wedge an agentic development environment into PostHog would have made both products worse.

Plus, we want Twig to work with other analytics tools too - not just PostHog. Keeping them separate gives us that flexibility.

So yes, we're still building PostHog. And now we're building Twig too. The future of product development is autonomous, and we want to be there first.

  • the twig team