Entire Launches Distributed Git Network for the Agent Era

Five months after the largest seed round in developer-tools history, Entire is reimagining version control for agent-scale development, allowing developers to mirror GitHub repositories on a fast, independent Git network built for agents.

July 8, 2026 — Today, Entire launched the preview of its distributed Git network, letting developers host repositories around the world instead of on a single central provider. Open under waitlist today, with active regions in the US, EU, and Australia, the preview lets developers mirror an existing GitHub repository onto Entire in one step: their code stays where it is, while their agents clone and pull from a regional Entire mirror. This offloads heavy, concurrent read traffic so agents can build without rate limits. In the coming months, Entire will let developers host new public and private repositories natively and plans to fully decentralize the network, enabling data residency, sovereignty, and scale in any region.

Entire has rebuilt the Git backend to handle simultaneous, high-volume agent activity, and initial testing has shown strong results, up to 25x over competitors such as Cursor Origin.

• Git Clone — how a repo gets copied down so a developer or agent can start working. Entire sustained ~570,000 clones/hour from a single repository.
  • 200 simulated clients shallow-cloning from Frankfurt (40%), Paris, London, and Dublin over ~3 minutes.
• Git Push — how changes get sent to the shared repo, so everyone (and every agent) builds on them. Entire sustained 586 pushes/second (or ~2.1 million/hour) to a single repository or branch.
  • 128 simulated agents pushing 1–10 files (2 KB each) per push over 2 minutes, each to its own branch. Tested on Entire native repositories.
• Clone + Push (mixed) — the real-world loop of pulling a repo down and pushing changes back, over and over, the way an agent actually runs. Entire sustained ~470 clone + push operations/second on a single repository.
  • 128 simulated agents running shallow clone → 5 pushes → repeat, at ~50–60 ms p50 latency.

In addition to these benchmarks and the underlying benchmark suite, Entire will open source its Git backend.

“By design, Git was always meant to be distributed. As Linus Torvalds put it in his 2007 Google Tech Talk: ‘If you’re not distributed, you’re not worth using.’ In the era of agents, centralized Git hosting has become a fundamental constraint, as the strain of billions of agents and developers hammering a central server shows up in the form of rate limits, high latency, or even outages,” said Entire CEO Thomas Dohmke. “Today, we begin to return Git to its original promise, with a distributed, and soon fully decentralized and open-sourced network of interconnected nodes around the world. By doing so, we enable any developer or agent to host their code in-region, pushing, pulling, and cloning close to where they operate, fast and without bottlenecks, while still part of a global, collaborative network.”

Since its launch in February, Entire has offered a semantic memory layer that helps agents stop repeating mistakes and lets developers see how and why any piece of software was made. It now integrates with every major coding agent, including Claude Code, Codex, Cursor, Factory AI, and GitHub Copilot, to automatically store each session, prompt, tool call, and checkpoint directly in the repository alongside the code. Entire announced new capabilities that turn that history into answers:

• Entire Blame — git blame shows who last touched a line of code; Entire Blame shows why: the agent session, prompt, and decision behind it.
• Entire Review — send a branch to multiple agents in parallel and get an intent-aware review grounded in the sessions, checkpoints, and diff.
• Code and Semantic Search — developers and agents can search through the entire history of not only their code changes, but why it was written.

“Session logs are now the second most important artifact in software development, and they belong in the repository alongside the code,” explained Dohmke. “With this semantic memory layer tied to the repo, agents stop repeating mistakes, which means higher accuracy, more productivity, and lower token spend. Developers can understand and verify what was built and why, which makes review far faster. And it opens up the possibility to build a new developer lifecycle, one that lets us understand and reason over the massive volumes of code AI agents now generate.”

Entire has grown to over 40 people across a global team located in the US, Australia, Germany, Spain, India, New Zealand, the UK, Czech Republic, and the Netherlands, with plans to expand headcount to 60 people by the end of the year as it builds the world’s next developer platform.