Former GitHub CEO launches Entire to solve AI coding agents' strain on centralized platforms

Former GitHub CEO Tackles Infrastructure Crisis in AI-Driven Development

Thomas Dohmke, former GitHub CEO who led the Microsoft-owned platform for four years, has launched a preview of Entire's distributed Git network designed specifically for the era of AI coding agents 1

1

2

2

. The platform addresses a growing problem: centralized hosting platforms are buckling under the weight of billions of concurrent requests from AI agents attempting to clone, pull, and push code simultaneously. Dohmke left GitHub in August 2025 to co-found Entire with Cole Driver, a former GitHub deputy chief of staff, and the startup has already grown to more than 40 employees across nine countries 4

4

.

The timing reflects an urgent need. GitHub itself froze new Copilot sign-ups as agentic usage broke its economics, revealing the strain on centralized platforms that weren't built to accommodate the explosive growth of AI-driven development 3

3

. Dohmke argues that while Git was designed to be distributed from day one, hosting platforms have routed developers into centralized systems that are now showing cracks under agent workloads .

How Entire's Distributed Approach Works

The preview, currently available by waitlist in the U.S., EU, and Australia, allows developers to mirror GitHub repos onto Entire in a single step 2

2

. This creates a parallel infrastructure where AI coding agents can clone and pull from regional Entire mirrors instead of hammering GitHub's central servers. The original code remains on GitHub, but the heavy read traffic gets offloaded to closer, faster copies designed to handle the unique demands of vibe coding workflows 3

3

.

Entire claims significant performance metrics compared to competitors. The platform can handle 2.1 million pushes per hour and 570,000 clone operations per hour, compared to Cursor Origin's 81,000 pushes per hour and 296,000 clones per hour 1

1

. In mixed tests designed to mirror actual agent behavior—cloning a repository, pushing changes, then repeating—Entire sustained roughly 470 operations per second with median latency between 50 to 60 milliseconds 3

3

. The company plans to open-source both the Git backend and benchmark suite, allowing independent verification of these figures 4

4

.

Beyond Mirroring: Agent Auditing and Semantic Memory

Entire's ambitions extend beyond solving rate limits and latency issues. The platform includes a semantic memory layer that captures every agent session, prompt, and tool call, storing them in the repository alongside the code itself 4

4

. This addresses a critical gap in current workflows: understanding not just what changed, but why an agent made specific decisions.

New features build on this foundation. Entire Blame traces a line of code back to the specific agent conversation that produced it, while Entire Review runs automated code reviews using that stored context 2

2

. A code and semantic search feature lets developers query the reasoning behind code changes, not just the changes themselves 3

3

. Dohmke argues that session logs have become "the second most important artifact in software development" and belong in repositories alongside code 2

2

.

The platform now integrates with every major coding agent, including Claude Code, Codex, Cursor, Factory AI, and GitHub Copilot 2

2

4

4

. This complementary positioning—working alongside existing tools rather than replacing them—is viewed as a key factor in the company's prospects, according to Tim Porter, a Madrona managing director who helped lead the firm's investment 2

2

.

Business Model and Future Plans

Entire launched in February 2026 with a $60 million seed round led by Felicis, with participation from Madrona, Microsoft's M12, and Basis Set Ventures, plus individual investors including Yahoo co-founder Jerry Yang and Y Combinator CEO Garry Tan 2

2

. Felicis called it the largest seed investment ever for a developer tools startup, valuing Entire at $300 million 2

2

4

4

.

The company plans to introduce commercial and individual pricing tiers after the preview period, mixing seat-based and consumption-based models alongside a free tier and open-source components 2

2

. The business model hints at monetization through Git hosting, agent auditing services, and potentially the data exhaust relevant to companies managing software agents 1

1

.

In coming months, Entire will enable native repository hosting for new public and private repositories, not just mirroring existing ones 2

2

3

3

. The longer-term vision involves full decentralization through interconnected nodes that let teams keep code in-region for data residency and sovereignty requirements while remaining part of a global system 3

3

4

4

. The company also plans to open-source its Git network and allow self-hosting, following through on Dohmke's stated goal of building "an open, decentralized, independent developer ecosystem for any agent and any human" 1

1

.

The participation of M12, Microsoft's venture arm, signals a cooperative rather than adversarial relationship with GitHub, at least for now 2

2

. However, Madrona described GitHub as "quickly becoming a legacy platform" in an earlier blog post, suggesting Entire's goal is "not only to supersede GitHub, but to superset it" 2

2

. The company expects to reach 60 employees by year's end as it scales operations 3

3

4

4

.