Fundraise

Meticulous raises $15M to test AI-generated code before it ships

What's the deal? London-based Meticulous has raised $15 million in Series A funding to help developers trust AI-generated code. Ethan Kurzweil at San Francisco-based ChemistryDealroom has a profile for this one. Try Dealroom → led the round, joined by Menlo Ventures.

Who else backed it? A long list of angels invested, including former Stripe operator Lachy Groom, Poolside cofounder Jason WarnerDealroom has a profile for this one. Try Dealroom →, Dropbox cofounder Arash FerdowsiDealroom has a profile for this one. Try Dealroom →, Vercel founder Guillermo Rauch, and Hex chief technology officer Caitlin Colgrove.

What does it do? Founded by brothers Gabriel and Quentin Spencer-Harper, Meticulous analyses a codebase and builds a checklist of every edge case — the rare scenarios that expose bugs or crashes. Whenever code changes, it simulates user flows before and after to flag what breaks.

Why now? AI agents are writing code faster than review and testing teams can keep up, slowing developers with long review cycles and more bugs. Meticulous gives developers a report of thumbnails showing each affected edge case before a change is merged.

"In a couple of minutes or less you understand the full impact of the change — you see it before your eyes — and can therefore ship almost as fast as your AI can write code," Quentin says.

What's the traction? Meticulous has grown annual recurring revenue fivefold in the past year, counting Notion, ElevenLabs, Dropbox, Wiz, and LaunchDarkly among its customers. No change ships without human input.

What's the endgame? The company currently tests frontend code only but plans to move into full-stack testing. Quentin spent more than 10 years at Palantir; Gabriel was an engineer at Dropbox.

What could go wrong? The founders concede Meticulous covers frontend code for now, leaving gaps as customers scale to millions of lines and thousands of edge cases across the full stack.

The signal: As coding agents flood teams with output, the bottleneck shifts from writing code to verifying it. Investors are betting the tools that keep pace with AI-generated code — not just the ones producing it — are the next layer of the developer stack.

Read more: Sifted

Image credit: Meticulous

More top stories