On Monday, former GitHub chief executive Thomas Dohmke unveiled Entire, a Git forge built for teams whose AI agents write code faster than people can review it. According to Entire, the platform replaces the pull request with Trails, dashboards of agent-scored evaluations on every code push, and runs on a new Git-compatible hosting system the company says pushes code up to 89 times faster than GitHub. The launch comes two days before GitHub’s annual Universe conference, and pitches Entire as a direct alternative to the company Dohmke used to run. Entire is available to any developer today, with hosting live in the United States, Europe, India and Australia.
Trails Replace The Pull Request
The biggest change is Trails, Entire’s replacement for the pull request. Dohmke wrote in a blog post on Monday that the pull request “has become an audit log that only the agent reads.” Instead of reading through every changed line, developers see a dashboard of scores from AI agents evaluating the code on each update, covering quality, security and what Entire calls slop.
Each agent runs independently in a cloud sandbox and submits a score and a rationale, according to Entire. A Trail can begin with a bug report, a feature request or straight from an agent, and can span multiple repositories. When the scores stabilise and all findings are cleared, the code can merge. Teams can also require a person’s sign-off or passing automated tests before anything merges. The shift sits alongside how developers code with AI today.
Hosting Built For Agent Traffic
Entire is built on a new source control server compatible with Git, backed by a distributed object store and a transactional ref database. According to Entire’s open-source benchmarks, its system pushes code up to 89 times faster than GitHub and 6 times faster than Cursor Origin, with 418 pushes per second against GitHub’s 4.7 in the company’s table. Teams can host repositories natively on Entire, replicate them across regions for data residency, or keep GitHub as the source of truth and mirror from it.
At launch Entire partners with Buildkite and Depot for continuous integration, and developers can migrate GitHub Actions workflows rather than rewriting them. Public repositories and self-hosting are on the roadmap, according to GeekWire. GeekWire reports GitHub suffered several significant outages in recent months, including one in August that lasted nearly eight hours, while developers and agents made 7.38 billion commits there in September alone.
Continuous Memory Carries Context Forward
The third piece is Continuous Memory, Entire’s attempt to stop agents re-learning a codebase from scratch in every session. Two open-source, local-first tools in the Entire command line make it possible: entire-graph, which maps how code fits together, and entire-brain, which indexes past agent sessions and attaches them to that map. Entire says the map saved up to 71% on tokens in its testing, by giving agents a precomputed picture of the codebase instead of making them grep through it.
Entire-brain can also connect to the Entire backend, so a team’s agent memories live in one shared space instead of on individual machines, and every Trail shows the memories the agent created. Reviewers thus see not just what changed, but why: the intent behind the logic, fixes already shipped, and what broke last time.
Why It Matters
Entire is betting that code review, not code generation, is now the bottleneck in software development. Dohmke said at a Madrona event last month: “Code review was always a bottleneck. I’ve never worked in a software company where I didn’t wait for somebody else to review my pull request, and you have to ping people in Slack three times. And there was this one guy that had to review everything because he knew what to look for. And now that bottleneck is 10 times bigger.” If agents keep multiplying the volume of code, a review system built for human-paced pull requests will either slow teams down or get skipped, and Entire’s Trails are an attempt to make review scale with generation.
Entire is not alone in rethinking code hosting for agents. Cursor launched its own code-hosting service, Origin, in August, and Anthropic has described building its own source control system as the number of code changes it merged each day grew from about 2,500 in May to more than 30,000 in September, according to GeekWire. The contest now is over where agent-written code lives and how it gets checked, a question at the heart of the wider AI coding market.
Background
Dohmke, who is based in Bellevue, Washington, left GitHub last year after nearly four years as chief executive. Entire now has 50 employees and is fully remote, and Microsoft’s venture arm M12 was among the backers of its $60 million seed round in February, led by Felicis with Madrona and Basis Set Ventures. That July, Entire launched by mirroring GitHub repositories on a faster network; Monday’s launch goes further, letting developers host code on Entire itself instead. Microsoft acquired GitHub for $7.5 billion in 2018 and has since folded it into its CoreAI division.
Source: Entire: Entire Git forge launch; GeekWire: Ex-GitHub CEO launches Entire
0 Comments