# Cursor Origin Puts AI Code Hosting Next to GitHub > Cursor launched Origin as a code-hosting beta for paid users while GitHub was still working through a broad service incident. - Content type: NewsArticle - Section: News - Published: 2026-08-17T20:48:00.000Z - Publisher: Arkolith Newsroom - Canonical HTML: https://arkolith.com/news/news-cursor-origin-github-outage - Plain Markdown: https://arkolith.com/news/news-cursor-origin-github-outage.md - Topics: Cursor, GitHub, AI coding, Developer infrastructure, Code hosting ## Article Cursor launched Origin, a code-hosting beta for paid users, on August 17 while GitHub was still reporting a broad service incident across developer workflows. The timing turned a product launch into a live question about where AI coding tools stop being editors and start becoming software infrastructure. The important boundary is narrow. Cursor is not saying every GitHub repo should move today. It says Origin can host repos, mirror GitHub projects, handle pull requests, browse code and put agents in the same place as the code they change. ## What Cursor launched Cursor's [Cursor Origin Code Hosting changelog](https://cursor.com/changelog/origin-code-hosting) says Origin began rolling out on August 17 in early beta on all paid plans. The first release covers repos, pull requests, code browsing and GitHub sync, with agent-native features still to come. The product makes Cursor more than an interface around code that lives elsewhere. In Cursor's [Cursor Origin documentation](https://cursor.com/docs/origin), the company describes Origin as a git forge for storing and sharing code. The beta can create repositories, push and pull with standard git, mirror GitHub repositories, open and merge pull requests, browse and search code, connect Vercel, Depot and Buildkite, and connect cloud agents to Origin repos. That list matters because [AI coding agents](/news/news-claude-code-auto-mode-default) create a review problem as much as a generation problem. If code, pull requests, checks and agents sit in separate systems, every handoff adds latency and policy risk. ## GitHub stayed the source of truth for synced repos Cursor's launch copy does not frame Origin as a hard GitHub replacement for existing projects. For synced repositories, the changelog says pushes keep going to GitHub, which stays the source of truth for work that started there. Pull-request comments and reactions can sync in both directions. That is the migration compromise. Teams can test Cursor-side browsing, review and agent work without immediately moving the canonical repo. It also leaves hard questions for regulated or security-sensitive organizations: who can view synced code in Cursor, which admins can disable Origin, and whether older privacy-mode teams need to change settings before using it. Cursor's docs answer part of that. Origin is available on Pro, Teams and Enterprise, not free plans. Admins can disable it for a team, and teams on legacy privacy mode cannot enable Origin until they switch to the newer privacy mode. ![Illustration: three blank folders and a brass switch on full-bleed dark paper](/article-images/news-cursor-origin-github-outage-1.webp) ## Why the GitHub outage made it louder GitHub's own [GitHub August 17 incident page](https://www.githubstatus.com/incidents/zkxwbgr0cnmx) shows the backdrop. GitHub began investigating impacted performance at 13:40 UTC and later listed affected areas including Git Operations, Webhooks, API Requests, Issues, Pull Requests, Actions, Pages and Copilot. At 14:04 UTC, GitHub reported about 20 percent error rates across numerous experiences and about 50 percent error rates for archive and raw repository content downloads. By 20:08 UTC, the incident was still not fully closed: GitHub said it was investigating sporadic Copilot authentication failures in some applications. The outage does not prove demand for Origin. It does prove the risk Cursor is trying to sell against. When code hosting, pull requests, CI, deployments and [AI coding](/news/news-openai-outage-codex-reset) all depend on one hub, a service incident becomes a delivery-chain event. ## What comes next The next test is enterprise behavior, not launch-day attention. Origin has to prove that mirroring, access controls, CI integrations and pull-request sync are reliable enough for teams that already have GitHub workflows. The sharper question is whether AI coding changes the code-hosting market itself. If agents are producing more branches and pull requests, the winning system may be the one that reviews agent work fastest while keeping humans in control. For now, Origin is an early beta with GitHub sync and admin escape hatches. That makes the launch important, but it also keeps the burden of proof on Cursor. *This article is informational only and is not investment, legal, tax or accounting advice.* ## Citation and verification - Cite the canonical HTML article: https://arkolith.com/news/news-cursor-origin-github-outage - Follow the links in the article to verify claims against the named primary or trustworthy sources. - Publication and update times above are the newsroom's recorded timestamps.