Google Labs has introduced Stitch Loop, a long-running product agent that connects to your codebase and live metrics to keep improving what you have shipped. Where coding agents stop once a pull request merges, Stitch Loop keeps working around the clock — patching vulnerabilities, speeding up checkout, diagnosing signup drop-offs — until your goals are met.

You start by connecting GitHub and your product context — logs, user behavior, feedback — through more than 180 integrations, custom MCP servers or local CLI uploads. Grounded in that context, Stitch Loop runs evergreen loops toward the measurable outcomes you care about, with three default loops out of the box and room to add your own. Each loop runs through four steps: Suggestions, surfacing opportunities grounded in evidence from your product; Solutions, each with tradeoffs and scoping, plus visual design options where relevant; Build, where Stitch Loop writes the code in an isolated cloud sandbox and opens a GitHub pull request for your approval, with screenshots and videos where relevant; and Reflect, monitoring live product data as it works toward your goal and learning from every result. Nothing merges until you review the pull request.

Built-in loops cover cybersecurity, latency and accessibility, plus custom loops such as improving checkout conversion. The security loop pairs Gemini 4 Carbon's cybersecurity capabilities with CodeMender, Google DeepMind's autonomous agent for code security — together they scan your codebase for security vulnerabilities before they become incidents, flagging each issue with evidence and opening a pull request with a proposed fix.

Stitch Loop also works as a service layer for coding agents such as Google Antigravity, and in Slack — use @stitch for daily insight digests and to create new loops — while the Stitch CLI and MCP let any coding agent pull persistent product context. The waitlist opens today, globally, with eligible Google AI Ultra users getting priority access at stitch.google.com/loop/waitlist.