Sutro bet GPT-4 plus rule-based compilers could replace your whole product team

An AI platform that turns a prompt into a production-ready app for web, iOS and Android, with hosting, maintenance and upgrades included.

That LLMs plus proprietary rule-based compilers can stand in for an entire product team — PM, design, engineering, hosting and scaling — not just drag-and-drop no-code.

$2.2M raised from Eniac Ventures and angels including OpenAI VP of Product Peter Welinder.

Tomas Halgas, who sold group-chat app Sphere to Twitter, and ex-Google and Facebook PM Owen Campbell-Moore founded Sutro in late 2021, planning for the transformative AI they were sure was coming.

The pair alternated running the company while Campbell-Moore took a day job at OpenAI; with Halgas back from Twitter as CEO, Sutro combined GPT-4 and other LLMs with proprietary compilers to generate full apps from a prompt, added a studio toolkit for visual and functional edits, and allowed custom code and integrations. It opened a waitlist in November 2023 with pricing still being set privately with vetted customers, pitched as 'a fraction' of hiring a dev team.

The compilers are the hedge: AI handles generation, but deterministic compilation makes the output production-ready rather than a demo.

It targets the years of undifferentiated plumbing — infra, auth, security — that every app repeats before the idea appears.

Hosting and maintenance in the subscription align Sutro's revenue with the app's whole life, not just the build.

High-level edits ('add ratings to posts') test whether the platform really understands product intent, not just code.

Position against the nearest category, not the obvious one: Sutro refused 'no-code' and sold the entire product team, because drag-and-drop still leaves the work with you.

As of 8 November 2023, Sutro had published its new landing page, opened access requests, and was finalising pricing privately with vetted customers; its team of roughly a dozen split across London and San Francisco.