

Thomas Dohmke, who led GitHub until 2025, launched Entire, a developer platform built for the age of AI coding agents. The startup positions itself as a direct rival to GitHub, the company Dohmke once ran.
Entire offers the core tools developers expect from a Git platform: repository hosting, code review, version control and build automation. But its central premise is different. It assumes AI agents will create much of the code, while humans will supervise, approve and merge changes.
Entire’s biggest departure is its replacement of the traditional pull request with Trails. In the current model, developers review a diff, leave comments and approve changes line by line. Entire argues that this process cannot keep pace when agents submit code continuously.
Trails gives developers a dashboard-style view of every code push. Multiple AI agents assess the change against custom evaluation criteria and assign scores. Developers can then decide whether to merge, request fixes or block the update.
The platform is now available to developers. Entire has raised USD 60 million in seed funding and says its infrastructure can handle the surge in AI-generated commits that is straining conventional developer tools.
Teams can define their own agents and choose models from Claude or OpenAI. Entire still allows humans to require manual approval or passing tests before a merge, preserving a layer of control over automated development.
The platform also introduces Continuous Memory, which stores agent sessions, prompts and project context within the repository. Entire says this helps future agents understand prior decisions and reduces repeated work.
For Dohmke, the move is also personal. He is now competing against the platform he once led, betting that the next phase of software development needs infrastructure designed for machines as much as for people
Also Read: Liang Wenfeng: How DeepSeek’s Founder is Disrupting AI with Efficiency