Automated bug killer

gm legends, happy Tuesday.

Here’s today’s lineup: Gamma API lets developers generate decks, reports, and carousels directly inside their own apps with a simple call; Onyx is an open-source AI teammate that connects to your Slack, Drive, Notion, and more so you can search and spin up agents without handing data to someone else’s cloud; Atla hunts down where your agents are failing, groups the errors, and helps you patch before users ever see them.

P.S. Building something new? Tell us about it → editorial@producthunt.co 🫶

Gamma API lets developers generate presentations, reports, and carousels inside their own products. Feed it text, transcripts, or structured data, and it returns polished output with layouts and themes already handled. Instead of bouncing between slide editors, the API turns raw input into content that’s ready to share.

🔥 Our Take: Slide decks have always been the tax you pay for having ideas. Gamma API doesn’t kill that tax, but it makes it trivial for apps to cover it on your behalf. That means fewer nights dragging text boxes around and more tools quietly spitting out the decks for you.

Onyx is an open-source AI teammate for work. Connect any LLM, plug in your Slack, Drive, Notion, GitHub, and more, then search across it all and the web in one place. You can spin up custom agents with MCP actions, do deep research, and use it on the web, in Chrome, or inside Slack. Self-host it to keep data inside your walls.

🔥 Our Take: Most “AI at work” tools forget everything that matters. The interesting swing here is letting you bring your own model, point it at your real sources with permissions intact, and give it hands via MCP when you want it to act. Open source means you can stop pasting private docs into random chats and just run it yourself.

Atla is built to stress-test AI agents before your users do. It runs through workflows, catches errors, groups similar failures, and ranks what’s worth fixing first. You can drill into traces, generate patches, and re-run tests until the agent behaves.

🔥 Our Take: Everyone’s hyped on shipping their next big agent, but most of them break the second they leave a demo. Atla steps in before that happens. It’s less “make your AI smarter” and more “stop embarrassing yourself in front of users.”

Justin Tahara asks which cities outside of San Francisco are becoming big startup hubs. He’s noticed a lot of teams and offices popping up elsewhere and wants to know where others think the next centers of gravity will be.

Some folks mention Vancouver, Toronto, Austin, and parts of Europe like Scandinavia as up-and-coming. Others say smaller cities are starting to shine because talent is cheaper, living is less brutal, and there’s less legacy baggage dragging things down.

What really comes through in the thread is this trade-off: you might lose some of the prestige or connectivity you get in old hubs, but you gain breathing room, affordability, and the chance to build without constant overhead.