Entire launches a Git forge for agent-human collaboration

Entire says it introduced what it calls the first Git forge built for agents and humans to work together.[1] The company describes the service as a GitHub alternative aimed at helping teams merge changes as quickly as coding agents create them.[1] Its Trails feature offers a heads-up display with agent-scored evaluations on every code push.[1] Entire says the review agents run independently in cloud sandboxes and submit both a score and a rationale.[2] The company says developers can also use Trails on top of pull requests.[2]

Entire says agents can address review findings in a loop while humans validate their progress.[1] Developers can merge without reviewing every line once all evaluations are satisfied, according to the company.[1] That does not mean teams have to remove human approval or CI from their merge process: Entire says they can configure merge requirements that include both human approval and passing CI checks.[2] For teams assessing the workflow, the distinction matters: agent evaluations can contribute to review, while teams choose the conditions that must be met before a merge.[1][2]

Entire says its source-control server is compatible with Git, and that private repositories can be hosted in regions including the United States, the EU, India and Australia.[1] Organizations can replicate projects and repositories across regions or keep them in a chosen region for data-residency needs, the company says.[2] Entire claims Git pushes are up to 89 times faster than GitHub and six times faster than Cursor Origin.[1] Separately, Analytics India Magazine reports Entire’s benchmark as 418 pushes per second, compared with 4.7 for GitHub.[3] Entire says it is partnering with Buildkite and Depot on build and test pipelines, and that existing GitHub Actions workflows can be converted using Buildkite’s `bk pipeline convert` or Depot CI.[1]

Entire says Continuous Memory syncs agent sessions into Git repositories so coding agents can search that data, and claims it saves up to 71% on tokens.[1] Its agentic-brain component turns previous sessions and prompts into durable memories that agents can query, according to the company.[1] During the preview, Entire says it charges only for agent usage.[4] After users use all their credit, they can purchase more and continue using agents at model-provider rates.[4] The company says it raised $60 million in February and describes the round as the largest seed round in developer-tools history.[1] Entire also says its team has grown to 50 people, with experience at companies including GitHub, Microsoft and Atlassian.[1]