Entire Launches Distributed Git Network for AI Coding Agents
Entire unveils a distributed Git mirror network sitting in front of GitHub, hitting 570K clones/hour and 586 pushes/sec, with the backend going open source.
Git's original design assumed a handful of humans running git push occasionally, not thousands of coding agents hammering the same repository with shallow-clone-then-push loops. Entire has launched a preview of a distributed Git network built to absorb exactly that load, placing regional mirrors in the US, EU, and Australia in front of GitHub while leaving the origin repo, CI, webhooks, and access controls untouched.
Initial benchmarks on a single repository are notable: roughly 570,000 shallow clones per hour, 586 pushes per second (about 2.1 million per hour), and a sustained 470 clone-plus-push operations per second at 50-60ms p50 latency. That mixed-loop number is the most telling one, since it mirrors how an actual coding agent behaves — pull, edit, push, repeat — a pattern traditional Git hosting was never built to handle at scale.
Adopting it requires minimal change: pointing an existing git remote at the nearest regional mirror works with tools like Cursor, Claude Code, Aider, or custom CI runners. Entire also plans to open source the underlying Git backend in the coming months, a move toward a fully decentralized model where organizations can self-host their own region for data residency needs.
The rollout still has real constraints — it's waitlist-gated, region coverage excludes APAC and South America, and the backend isn't open yet — but it represents the first concrete, benchmarked attempt to solve Git's scalability problem for the agent era.