Propolis: AI swarms that eliminate QA by testing like real users

Hey Product Hunt! Marc and Matt here, co-founders of Propolis. Propolis was born from a series of frustrating moments: watching the smartest engineers on our teams waste time writing tests or getting Operator (failed miserably) to test for them, only to watch them break with every UI change. At Palantir, Airtable, and various startups - we'd seen this pattern too many times. Our "aha" moment: what if AI agents could test products like humans do, by building an independent understanding of how to use the product? That's Propolis. Our autonomous browser agents: - Learn how to use your application - Test every possible user flow - Adapt instantly when your product changes - Deliver detailed bug reports with reproduction steps While the first practical application is QA, we know this can go so much further. Simulating synthetic users that can launch before putting a new feature out into the world, providing feedback on what pieces to optimize. We're also recreational beekeepers (hence the name Propolis—the protective resin bees use to shield their hives, just as our swarm protects your product from bugs). As a special PH launch offer, the first 10 companies who sign on to use Propolis get some of our recently harvested honey! 🍯 Marc & Matt

Marc Papazian

Propolis

Maker

📌

29d ago

Evgenii Zaitsev

Propolis sounds like a real breakthrough in QA! Autonomous AI agents that test like real users and adapt to UI changes could save teams so much time and frustration.

How do these agents handle complex workflows or edge cases that might require domain-specific knowledge?

Also, the idea of using synthetic users for pre-launch feedback sounds promising. Any plans to expand into that space soon?

Supa Liu

Propolis takes a novel approach to QA—AI swarms that behave like real users, uncovering bugs across product flows with zero manual effort.

Propolis

AI swarms that eliminate QA by testing like real users

Do you use Propolis?

Review Propolis?