Cursor Unveils Origin, a GitHub competitor!
Cursor just dropped "Origin," a GitHub competitor.
Cursor is moving way beyond "VS Code with AI." They just teased Origin, a Git-compatible forge whose landing page describes it as a tool built for "the agentic era," with code "moving faster than any infrastructure was built to handle." It's currently waitlist-only.
According to developer Mark Kretschmann, Origin is designed around the assumption that swarms of AI agents will be cloning, branching, committing, rebasing, and reviewing code in parallel, something traditional GitHub workflows weren't built for.
He also connects it to Cursor's recent Graphite acquisition, which brings stacked PRs, review flow, and merge queues. Combine that with Origin's hosting layer, and Cursor starts looking less like an editor and more like an end-to-end "AI software factory."
My take: this feels inevitable rather than surprising. If agents are about to generate far more code than humans can review, the bottleneck shifts to merging and trust, and someone will have to build infrastructure to fill that gap.
Replies
This reminds me a lot of what Graphite has been pushing with stacked PRs and streamlined reviews. Combining that philosophy with a Git hosting platform could be a powerful combination .
That will 100% be the bottleneck. Curious how fast GitHub will respond.
would this be a direct competition or a duopoly with Github? Not sure if agentic code and human-written code are 100% direct competitors. What about people who still prefer (mostly) human-written code because of trust and responsiblity?
Bond
super pumped to see how they 10x github
the bottleneck rohan points to is real, but i'd push it one layer deeper. it's not just merging and trust in the repo, it's verifying that what the agents actually did at runtime was correct. swarms cloning and committing in parallel means you also need infrastructure to validate the API calls, retries, and async state transitions those agents triggered, not just whether the PR looks clean. origin solves the forge layer. the observability layer for what agents do outside the repo is still wide open.
This feels like a logical expansion, not just a product launch.
If Cursor believes AI agents will generate and modify code at a much higher volume than humans, then owning only the editor was never going to be enough. The real bottleneck shifts upstream and downstream: branch management, review flow, merge safety, provenance, and trust.
That’s why Origin is interesting. It suggests Cursor wants to own not just code generation, but the entire system where AI-written code gets coordinated, reviewed, and shipped. Graphite gave them the review workflow layer. A forge gives them the hosting and control layer.
I also think Raj’s point is important: repo infrastructure is only one half of the problem. The next big category may be runtime verification and observability for agent actions outside the repo. Origin could solve the code movement layer, but the trust layer around execution still looks wide open.