Entire is a Git forge — a GitHub alternative — built for teams where AI coding agents write most of the code. It launched publicly on October 26, 2026, led by former GitHub CEO Thomas Dohmke. Its main ideas are Trails, which replace pull-request review with agent-scored evals, distributed Git hosting in the US, EU, India and Australia, and Continuous Memory, which stores agent sessions alongside the code.

What are Trails in Entire?

Trails are Entire's replacement for the pull request. Instead of a reviewer reading a diff, each push is scored by agents against custom evals such as security and code quality. Coding agents fix the findings in a loop, the developer watches the scores, and when all evals pass the change can be merged without reading every line. Results are available in the web app and through the open-source Entire CLI.

How much does Entire cost?

Entire did not publish a pricing page at launch. Launch coverage reported that hosting is free to start, new users get a $100 credit for Trails review agents, and review-agent tokens are then charged at market rates for the model you choose. Check entire.com before committing, as pricing may change.

Is Entire a replacement for GitHub?

For private repositories in AI-heavy teams, it is designed to be. It hosts code, replaces pull requests with Trails and connects to CI through Buildkite and Depot, with tools to convert GitHub Actions workflows. At launch it lacked public repositories and self-hosted Git nodes, which Entire lists as upcoming, so open-source projects and teams needing on-premise hosting should wait.

Quick Answer

What Is Entire? Thomas Dohmke's Git Forge for AI Agents

Published: October 27, 2026

The short answer

Entire is a Git forge built for AI coding agents and the humans who supervise them. Former GitHub CEO Thomas Dohmke’s company launched it to every developer on October 26, 2026, as an independent alternative to GitHub. It rests on three ideas:

Trails replace the pull request: agents score every push against your evals, and you merge when they are green.
Distributed Git hosting with pushes Entire says are up to 89x faster than GitHub, in US, EU, India and Australia regions.
Continuous Memory stores prompts, agent sessions and decisions next to the commits, so the next agent starts with context.

Entire raised a $60 million seed round in February 2026 and has a team of about 50.

What Entire includes at launch

Feature	What it does
Trails	Agent-scored evals (security, quality, risk) on every push; agents loop on findings; merge when evals pass
Code hosting	Git-compatible distributed server; regions in the US (Ohio, Oregon), EU (Frankfurt, Ireland), India (Mumbai), Australia
Session storage	Prompts, tool calls and decisions captured per commit by the Entire CLI; secrets redacted before storage
Continuous Memory	`entire-graph` (codebase graph, up to 71% fewer tokens per Entire) and `entire-brain` (memories from past sessions), both MIT
Agentic search	One query over code, commits and agent sessions
CI	Buildkite and Depot integrations; convert GitHub Actions workflows with `bk pipeline convert` or Depot CI

Entire says it “works with any agent”; its examples show Claude Code and Codex sessions attached to the same Trail.

Why the pull request is the target

Dohmke’s argument is that agents now write code faster than people can read diffs, so line-by-line review has become a bottleneck and the pull request mostly an audit log. Trails move review up a level: humans define what “good” means as evals, agents check every push against them, and humans validate the outcome and the trace of how it was produced. Whether teams will trust merging code they have not read is the open question; Trails keep the diff available, but the design assumes you will often not open it.

Speed and regions

Entire publishes pushes-per-second figures from its own tests: Entire 418, Cursor Origin 65, GitLab 10, GitHub 4.7 — hence “up to 89x faster than GitHub” and “6x faster than Cursor Origin”. These are vendor benchmarks; the relevant test for you is agent fleets cloning and pushing at once, which is the load Entire says it built for.

Pricing

No pricing page existed on October 27, 2026. Launch coverage reported a free start, a $100 credit for Trails review agents, and model tokens billed at market rates after that. Budget for review agents as you would for any LLM usage — they run on every push.

Entire vs GitHub vs Cursor Origin

GitHub remains the default for public open source, Actions, Packages and the broadest integrations.
Cursor Origin is the review-and-merge surface for teams living inside Cursor; see Cursor Origin vs GitHub vs GitLab.
Entire is the most opinionated bet on agent-first review, independent of any editor or model vendor.

Upcoming on Entire’s roadmap: self-hosted Git nodes, public repositories and live-streamed sessions. A low-risk way to try it is to mirror one active private repository and run Trails alongside your existing pull requests for a few weeks.

Last verified: October 27, 2026.