Former GitHub CEO Raises Record $60M for AI Coding Tool to Manage Agent-Written Code

Former GitHub CEO Launches Entire with Record Seed Funding

Thomas Dohmke has emerged from stealth with Entire, securing a $60 million seed round at a $300 million valuation in what Felicis, the lead investor, calls the largest-ever seed funding for a developer tool startup 1

1

. The former GitHub CEO, who stepped down in August 2025 after four years leading Microsoft's GitHub division, is tackling a pressing challenge in AI coding: how developers can effectively manage AI-written code as AI agents increasingly become the primary producers of software .

The Thomas Dohmke startup attracted heavyweight backing beyond Felicis, including Seattle-based Madrona, Microsoft's M12 venture arm, and Basis Set. Individual investors include Datadog CEO Olivier Pomel, Y Combinator CEO Garry Tan, former Yahoo CEO Jerry Yang, and developer community voices Gergely Orosz and Theo Browne . Dohmke's track record overseeing GitHub Copilot's growth—one of generative AI's earliest commercial successes—provided strong credibility for his vision of reimagining the software development lifecycle.

How Entire Addresses AI-Generated Code Management Challenges

Entire offers an open-source tool designed to help developers better manage code written by AI agents, addressing what Dohmke describes as an "agent boom" where massive volumes of code are generated faster than humans can reasonably understand 1

1

. The code management platform features three core components: a Git-compatible database to unify AI-produced code, a universal semantic reasoning layer enabling multiple AI agents to work together, and an AI-native user interface designed for agent-to-human collaboration 1

1

.

The first product Entire is releasing is the open-source tool Checkpoints, which automatically pairs every piece of software an agent submits with the context that created it, including AI prompts, reasoning, transcripts, and token usage details 5

5

. This approach allows human developers to review, search, and learn from why AI agents made specific decisions. "That way you have code and context together, and then it allows you to do things like identifying the intent of that code, of that change, or just figuring out when was that code generated and with what prompt," Dohmke explained .

Managing AI Coding Agents in an Increasingly Crowded Market

Entire enters a competitive landscape where AI coding has become one of generative AI's earliest success stories, with major players including OpenAI, Anthropic, Google, Microsoft, and Cursor offering their own platforms 4

4

. However, Dohmke positions Entire not as a direct competitor but as a complementary platform. "We are not training models or building agents, we are integrating with them," he told Axios. "Our platform will be open-source, independent, and scalable for every developer and agent to host their code and agent context" 4

4

.

Checkpoints launches with support for Anthropic's Claude Code and Google's Gemini CLI, with integration for OpenAI and GitHub expected within weeks . The company initially targets open-source developers, AI-native companies, and small startups, recognizing that popular open-source projects are particularly overwhelmed with suggested code contributions that may or may not be usable 1

1

.

Why This Matters for Software Development Teams

The platform addresses several critical pain points in modern development workflows. When AI agents make mistakes, developers often provide follow-up prompts with remediation guidance. Entire can save that guidance and make it available to a team's other AI agents, enabling them to avoid repeating errors 5

5

. This reuse of troubleshooting advice streamlines development and reduces unnecessary inference costs. Before tackling a technical difficulty, developers can review Entire to determine if a colleague already solved the issue in a past session, eliminating duplicate work.

S. Somasegar, managing director at Madrona who previously served as Microsoft's developer tools chief when the company acquired Dohmke's HockeyApp in 2014, emphasized the need for new infrastructure: "All the workflows that you imagined in your developer platform have to be reimagined. GitHub is a fantastic platform, but GitHub was built for an era where it was human beings doing all the work" . While acknowledging that GitHub and other AI companies are likely thinking about similar questions, Somasegar believes the market is large enough to support multiple players.

Entire operates as a fully remote company with 15 employees who previously built developer tools at GitHub and Atlassian, with plans to expand headcount as it works toward a broader platform launch later this year . The startup's approach of storing prompts alongside code in Git-compatible repositories represents a fundamental shift in how software teams might collaborate with AI agents, moving beyond simply saving AI-generated code changes to preserving the reasoning and decision-making context that produced them.