Former GitHub CEO launches Entire, an agent-first coding platform

Entire runs on the decentralized Git network the company announced in July. Entire says the network is designed to withstand the sharp increase in AI coding activity that has caused mounting GitHub outages.

The platform replaces pull requests with what Entire calls Trails. Entire describes Trails as a heads-up display, similar to one used by an airline pilot, that provides customized evals scored by agents on every code push. All data is available through Entire’s web interface and an open-source CLI, allowing coding agents to address findings in a loop while the human developer validates progress. When all evals are satisfied, developers can merge code without looking at the code itself, according to the company.

Dohmke said Entire treats the idea, the agent trace and the outcome of a developer’s sessions as first-class artifacts for review alongside the code. It uses multiple agents to generate evals that verify the quality and success of AI-generated code, he said, so developers can merge without reviewing every line in the diff.

Entire claims its Git hosting is up to 89 times faster than GitHub and six times faster than Cursor origin. The company says developers can create, migrate and host private repositories in multiple locations, starting with the US, India, Australia and the EU.

Another feature, Continuous Memory, keeps developer session logs next to code in the repository as a data source. Agents can search those logs for answers, which Entire says helps them use up to 71% fewer tokens. Entire said Continuous Memory consists of a codebase graph and an “agentic brain” that uses previous sessions and prompts as lessons so agents do not repeat mistakes.

Entire is available to try today for all developers, ZDNet reported.