AI coding agents have made generating software dramatically faster. Former GitHub CEO Thomas Dohmke believes the infrastructure around that code has not kept up. His startup Entire has now launched a developer platform built specifically for teams where humans and AI agents work on the same repositories.
Unveiled on October 26, Entire combines three pieces: a new approach to code review called Trails, distributed Git hosting and persistent memory for coding agents. Instead of positioning itself as another AI coding assistant, Entire is targeting everything that happens after an agent starts producing code — storing it, understanding why it changed, reviewing it and deciding whether it is safe to merge.
Trails is the biggest departure from traditional development workflows. Conventional pull requests ask developers to inspect code differences line by line. Entire instead lets teams define automated evaluations that run when new code is pushed. AI agents can inspect changes and surface issues, while human developers retain control over approvals and merging.
Entire is also launching native distributed Git hosting across regions including the US, Europe, India and Australia. The company claims its infrastructure can process pushes up to 89 times faster than GitHub and six times faster than Cursor Origin in its open-source benchmarks. Those figures are company-reported benchmarks and will vary with workload and infrastructure.
The third layer, Continuous Memory, stores previous agent sessions, prompts and development context alongside the repository so future agents can retrieve what happened before instead of repeatedly rebuilding context. Entire says its codebase-graph system reduced token usage by as much as 71% in internal testing.
Entire had already raised a $60 million seed round at a $300 million valuation in February, led by Felicis with participation from Madrona, M12, Basis Set and others. The October announcement is therefore a product launch, not another funding round.
Dohmke’s bet is that AI will make code abundant. If that happens, the valuable developer infrastructure may shift from helping people write more code to helping them understand, verify and manage everything the agents produce.
Filed by
Startup Unplugged
