Cursor Origin Release Date: Status Tracker (August 2026)

Yes, as an early beta. On August 17, 2026 Cursor's changelog entry "Origin Code Hosting" said Origin "begins rolling out today in early beta on all paid plans," and the product page at cursor.com/origin now reads "Early beta now available on all paid plans" where the waitlist form used to be. The docs at cursor.com/docs/origin narrow that to Pro, Teams and Enterprise, say it is not available on free plans, and warn that access opens in stages, so you may not see it the day it becomes available for your plan.

This page exists to save you the re-checking. The table below is verified against Cursor's own pages on the date in the caption, and that date is the thing to trust when you come back.

The ownership row is the odd one out in a release table, and it earns a line because the two events landed three days apart. Cursor's post "Cursor is now a part of SpaceX" went up on August 14, 2026, and the Origin beta opened on August 17, so the company hosting your code on Origin was already a SpaceX subsidiary on day one of the beta. That proximity proves nothing on its own. What the post promises is compute for training models; it says nothing about Origin's roadmap or pricing, so the deal terms belong in the due-diligence file rather than in a feature forecast.

It was beaten by a month, and it was never Cursor's number. Write-ups of the June Compile announcement settled on a fall window: one guide described Origin as "available this fall," and developer-news coverage called it "waitlist-only ahead of a fall 2026 launch." Those phrasings were repeated widely enough that fall read like an official date. Cursor never named a season, and the early beta landed on August 17.

The lesson transfers to the next date you will hear about, which is general availability. Cursor's own material is still silent on when the beta label comes off. The changelog's one forward-looking line is that agent-native features ship soon, with no quarter attached. Anything more specific than that is reporting until it appears on a cursor.com page.

The repetition is what makes this hard to see from outside. Each write-up is reasonable on its own terms, and none of them claims Cursor said it. Stack a few together and a season starts reading like something anyone could look up. The check is one click deep: follow the sentence back and see whether it lands on a cursor.com page or on another article.

None of that makes the reporting malicious. It just means a window is unsourced until it isn't, which is the thing to write down if a plan depends on it.

Through your existing paid plan, not a waitlist. Cursor's docs describe the setup as claiming a codebase name, which becomes the namespace every repo URL lives under, then creating a repo or syncing one from GitHub. The access guide covers the details and the gotchas; the short version fits in three steps.

1. 1Open cursor.com/codebase on a Pro, Teams or Enterprise account and select Get Started. On a team, any member can claim the codebase name; admins then grant repository access from codebase settings.
2. 2Choose the codebase name carefully. During the beta you cannot change it once claimed, and it is part of every repo URL (cursor.com/codebase/{owner}/{repo}).
3. 3Create a repo with New, or bring one over with Sync from GitHub, then install the Origin CLI (origin auth login) so git push and pull work against Origin remotes. Teams on legacy privacy mode have to switch to Privacy ModeCursor's setting that guarantees code data is not used for training by Cursor or its model providers, and that an admin can enforce org-wide; data-retention terms are a separate, contractual layer. Press Enter for the full definition. before any of this appears.

If you are on a paid plan and still do not see Origin, the docs' own explanation is staged rollout, and the other two causes are an enterprise admin who opted the org out and a team still on legacy privacy mode. Check those before assuming something is broken. Most people arrive at this question because somebody else asked them first, usually with a planning document open, and the answer I give now is that the beta is real, that access is a plan question rather than a queue, and that the review features the demo sold are the part still to come.

Products at this scale leak their next stage through infrastructure before anyone posts a date. Half of the signals this page used to watch for have now fired: documentation exists, the changelog carries Origin, and first-hand reports of access are ordinary. These are the ones left worth watching.

Those rows state a ranking without arguing for it. Pricing sits at the top because it is the slowest of the six to produce: somebody has to settle the unit, whether that is per seat or per repository, and the wording has to survive a legal review. That is not work anyone does speculatively. I should flag that I am reasoning from how launches generally go rather than from anything Cursor has published about this one.

My instinct is still to watch the pricing page, and it is still the wrong habit. Pricing is the loudest item on the list, and going by how fast the beta itself was written up, coverage will put it in front of you without your refreshing anything. Watch the docs instead: the settings pages can lose their redesign caveat quietly and sit that way a week before anybody mentions it.

Tracking got easier on August 17, because the docs now change when the status does. Cursor's changelog is the push channel, cursor.com/docs/origin is the pull one, and the rows above are what to compare between visits.

This is the living part of the page. Each revision adds a dated row, so a returning reader can tell in one glance whether anything moved since the last visit. Two months in, the honest summary is: the product shipped, and the interesting features did not yet.

That July row still does work, and it is the one most trackers leave out. A page that only updates when something happens looks exactly like a page nobody has opened since June, and a reader has no way to tell which one they are on. The August 22 row is the same idea in the other direction: it records what the beta did not bring, so the next visit can tell whether that list shrank.

If this page and something you read elsewhere disagree about the status, open cursor.com/docs/origin before you open anything else. The dull explanation is usually the right one: a cached copy of the June coverage, or a summary written from it and stamped with a later date. The docs settle it in a few seconds, and they are where the table above gets checked from anyway.