Is Cursor Origin Free? Free Tier and Private Repo Outlook

Not today, and the answer is now written down. Cursor's Origin docs say code storage is available on Pro, Teams and Enterprise plans and "not available on free plans." So the early beta that began rolling out on August 17, 2026 is free in one sense, it adds nothing to a paid plan's price, and paid-only in the sense that matters to a Hobby user: there is no way in without upgrading. A separate Origin price, in either direction, is still unpublished.

The snippet-shaped version, for anyone who wants the answer in one glance.

Yes, and that is the strongest reason to expect one from Origin eventually. A free tier is how a forge seeds adoption: developers try it on a side project, then carry it into work. A paid-only forge asks a team to pay before anyone has felt the workflow, and no major host has been willing to make that bet at general availability. Cursor has made it for the beta, which is a different thing: a paid-plan gate on an early beta reads as rollout control, not a pricing decision.

The four rows do not deserve equal weight. Graphite's is the one to read closely, since it is the same team pricing the same review layer, and its free line already sits at the boundary between a personal account and an organization. What does not carry over is the cost underneath. Graphite sells a layer that runs on somebody else's host, and Origin would be the host, which puts storage and bandwidth inside a free tier Graphite never had to fund.

Cursor's own products follow the same pattern. The editor's Hobby plan is free, with metered Agent requests and Tab completions rather than a time limit, and that Hobby plan is exactly the one the Origin beta excludes. None of this is a promise about Origin. It is the gravity of the market Origin launches into, and swimming against it at GA would be a genuinely surprising choice.

There is a limit to what pattern-matching buys you. Every host in that table ships a free tier, so "Origin will have one too" is a reasonable bet, and for a while that is where I would have stopped. It is the wrong place to stop: the bet worth making is about what sits inside the tier, since GitHub's free plan and Bitbucket's free plan both count as free tiers while drawing completely different lines around who and what they cover.

If a free tier exists, free private repos are table stakes. GitHub's Free plan includes unlimited private repositories, and GitLab and Bitbucket both cover private repos on their free tiers with a five-user cap. A forge launching in 2026 that charged for private repos would be behind every incumbent on day one, on the exact feature most working teams need first.

Where the limits would sit is the harder question. The incumbents draw the line in two different places: GitHub caps features (advanced review tooling, more CI minutes) while keeping repos and collaborators unlimited, and GitLab and Bitbucket cap people, five users each. The beta draws a third line, plan tier, and which line Origin draws at GA is unknowable until pricing ships.

Which of those two lines is worse depends on what your team is doing next. The table cannot rank them for you. A people cap fails at a single moment, when the sixth person joins and the plan stops working that afternoon. A feature cap fails slowly, as a review workflow nobody quite adopts because the part that would make it worth adopting sits one tier up. If you are hiring, the five-user caps bite first. For a team whose headcount is flat but whose review load is climbing, GitHub's shape is the more comfortable one.

The five-user caps also do not settle what counts as a user. Five is a clean number until there is a service account running CI and a contractor on for six weeks, and whether either of those consumes a seat is worth settling with the vendor before you plan around a headline cap.

This is the wrinkle that makes Origin's free tier harder to predict than a normal forge's. Traditional free tiers meter humans: seats and build minutes. Origin's whole pitch is hosting for code that agents write, and agents change what costs money. A team of five people with twenty agents in flight generates automated review, merge-queue work and conflict resolution continuously, machine work that a seat count never sees.

Cursor's existing free products already show the likely shape. The editor's Hobby plan meters "limited Agent requests," not hours in the editor, and Graphite's free Hobby plan limits AI reviews, not repositories. If Origin follows its siblings, a free tier would more plausibly cap agent activity, reviews run, merges queued, than repo or collaborator counts. That is inference from the same company's published plans, not an announcement.

For a large org the whole question is beside the point. A 200-engineer org is not going to standardize on a free plan, whatever its caps allow, so adoption there turns on the contract and the due-diligence list instead. Free tiers decide things for small teams and side projects, which is exactly where a five-user cap turns into an argument about who gets an account.

Not much, and that is the point: no decision you make today should depend on a price that does not exist. A few low-cost moves keep you positioned without betting on rumor.

1. 1If you are already on a paid plan, get Origin access and sync a non-critical repo; the beta costs nothing extra and GitHub stays the source of truth.
2. 2Keep using the free tier you already have. GitHub's free plan with unlimited private repos is the benchmark Origin has to beat anyway.
3. 3Write down what you currently pay your forge, per seat and per CI minute. That number is your comparison baseline the day Origin pricing ships.
4. 4Ignore any page quoting an Origin price before Cursor does. Today, all of them are guesses.

The third step is the one that gets skipped, and it is the only one with a number in it. When Origin pricing lands somebody will ask whether it is cheaper than what you have, and without last month's invoice split into seats and CI minutes, that conversation runs on impressions. Write the number down while nobody is waiting on it. The quieter trap is deferring a review-workflow fix until Origin ships, which buys you more months of the bottleneck you already have.