Entire Launches Distributed Git Network for AI Coding Agents
Entire, a developer-platform startup founded by former GitHub CEO Thomas Dohmke, has introduced a preview of a distributed Git network designed for the age of AI coding agents. The goal is to distribute code hosting across regions, mitigating the strain on central providers as agentic usage increases.
July 8, 2026 – 1:00 pm
As more coding agents clone and pull code simultaneously, central hosts struggle to keep up, a challenge GitHub experienced when it froze new Copilot sign-ups due to high demand. Entire’s solution is a mirror system. Developers can now mirror an existing GitHub repository with one step, allowing agents to clone and pull from a regional Entire copy instead of the original host.
This offloads heavy concurrent read traffic, ensuring agents can continue working without hitting rate limits. In the coming months, Entire plans to enable hosting of new public and private repositories directly on their platform. The long-term vision includes building a network of interconnected nodes, enabling data residency and sovereignty while maintaining global accessibility.
Dohmke argues that "by design, Git was always meant to be distributed," addressing the centralization constraints arising from billions of agents and developers accessing shared servers. He is not alone in this belief; he has also backed Tangled, a separate decentralized Git initiative.
To support its performance claims, Entire published benchmarks from internal testing, demonstrating sustained rates of 570,000 clones per hour, 586 pushes per second, and approximately 470 combined clone-and-push operations per second on a single repository. They claim median latency remained around 50 to 60 milliseconds when simulating agent activity.
The network is built upon the company’s existing semantic memory layer, launched in February, which aims to prevent agents from repeating past mistakes. This layer integrates with various AI coding tools like Claude Code, Codex, Cursor, Factory AI, and GitHub Copilot, storing session data alongside the code in the repository.
New features include Entire Blame, which provides insights into changes made by specific agent sessions, and Entire Revi, offering a detailed view of revision history.