On the morning of August 18, a new tab appeared in the Cursor editor: Codebase. After weeks of quiet, Cursor began rolling out Origin, its own code hosting platform, to paid users. Roughly three and a half hours later, GitHub's status page lit up with a global outage lasting 6 hours and 42 minutes — web and API error rates climbed to nearly 20%, and archive and raw file downloads hit 50%. A carefully planned platform launch collided with an uncontrolled outage at the incumbent, all on the same workday.
Origin is not another GitHub. It aims to put code, pull requests, and AI agents in the same window: developers can name repositories with independent URLs and push via the command line, while built-in hosting capabilities — storage, permissions, checks, and merging — work entirely without opening a browser. Agents act directly on files on screen, editing PRs in place and pushing branches, all inside the editor where the code is written. At launch, Origin already had three partners: Vercel generates preview deployments for every PR and ships to production on merge; Depot and Buildkite provide CI, both able to run existing GitHub Actions workflows as-is.
But the smartest move is Origin's decision to let GitHub remain the "data source." Once a GitHub organization is connected, existing repositories appear side by side with Origin-native repos, pushes still sync to GitHub, access permissions follow GitHub's existing read/write settings, and PR conversations sync bidirectionally — a comment on one side appears on the other within seconds. This "read-only mirror" approach avoids the high-risk path of full source-hosting migration: migrations involve CI, compliance evidence, audit records, branch protection rules, and every engineer's habits. Few teams would approve a full migration for an early beta. But if Origin's review experience is good enough, developers' actual work will gradually shift — and the data source follows attention.
While the timing was seen as opportunistic, the deeper story is that "agent-era high-frequency commit/clone traffic is overwhelming traditional Git infrastructure." Origin's agent-scale performance numbers: 296,000 clones per hour and 22 commits per second per repository. When a parallel agent workflow can generate hundreds of commits and pulls in hours, throughput and conflict resolution become bottlenecks — Origin builds parallel PR management, automatic merge-conflict resolution, and 400ms global latency in as native capabilities.
Core pitch: putting the agent in the same interface as code and PRs
Cursor didn't reinvent version control — it reinvented the interface between agents and version control. Developers can ask questions about files on screen, hand review comments to an agent to edit PRs in place, or have it push branches directly. Workflows that once bounced between the GitHub web UI, CI logs, code review comments, and the IDE are now compressed into one.
The "read-only mirror" is the decisive play
The real cost of migration isn't tooling — it's compliance evidence, audit records, branch protection rules, and engineer habits. By syncing to GitHub, inheriting permissions, and bidirectionally syncing PR conversations, Origin effectively outsources migration cost to attention migration: once developers habitually review diffs, comment, and merge PRs in Origin, the data source follows. It echoes how GitHub used "social coding" to marginalize SourceForge — except this time the social layer is agents.
Agent-era Git infrastructure bottlenecks are on the table
GitHub's 6h42min global outage won't be the last. As parallel agent workflows commit and clone at tens of times the usual frequency, throughput, concurrent merges, and Actions queuing on traditional platforms become crushing bottlenecks. Origin's hourly clone count (296,000) and per-second commit rate (22 per repo) are SLAs traditional hosting never promised — not a gimmick, but an agent-era engineering baseline.
What changes
Origin redraws the boundary of AI coding tools: Cursor moves from IDE to platform, turning "writing code" into "operating code." When agents take over code hosting, GitHub loses not just user engagement time but the entry point to developer workflows in the AI era. And GitHub's global outage that same day made the fragility of single-platform dependence concrete for every CTO.