Skip to content
Neural Daily

ToolsNews

Entire opens Git forge where agents run the review

🇭🇺 Olvasd el magyarul →

Former GitHub CEO Thomas Dohmke has opened Entire's Git platform to everyone. Trails replace pull requests, with review agents scoring evals on every push.

By: Mikola Máté Patrik · · 2 min read

Entire feature grid: 89x faster pushes than GitHub, Continuous Memory, Review with Confidence, 71% lower token spend, CI for Agentic Scale and 99.99% availability
Image: Entire · Press kit · Official image from the Entire launch blog post · Source

What launched?

Entire made its Git platform generally available on Monday (Entire). The company was founded by former GitHub CEO Thomas Dohmke, and the platform runs on the decentralized Git network it announced in July. ZDNET says the goal is to withstand the load from AI coding that has caused mounting GitHub outages (ZDNET).

Trails instead of pull requests

Instead of diff-based review, Trails give developers a cockpit-style display: review agents compute eval scores on every push. Three evals run by default (Security, Slop and Human Review Required), and custom evals can be added in the UI or as JSON. The agents run Claude Code and Codex in Daytona or E2B sandboxes. Findings get severities, and Gates can require approvals and CI before merging.

All data is also available through the open-source Entire CLI, so a coding agent can fix findings in a loop while a human checks progress. Entire says that once all evals pass, code can be merged without reading the diff.

Hosting and memory

The company claims Git hosting up to 89 times faster than GitHub and six times faster than Cursor Origin. Private repositories can currently be hosted in the US, India, Australia and the EU. Continuous Memory keeps developer session logs next to the code in the repo for agents to search, which Entire says cuts token use by up to 71 percent. Buildkite and Depot joined as CI partners.

What to keep in mind

The speed and token figures are the company's own claims. The direction is clear, though: when agents write most of the code, review shifts from reading line by line to automated checks and session history. Try it on a side project before moving a main repo.

    Produced with the help of AI, reviewed by an editor. How we make our articles

    Reviewed and edited by the Neural Daily desk. About

    Related articles