AI code editor whose cloud agents work on their own branch and open pull requests; Pro is $20 a month with extra usage billed at API rates.
Developers who want the agent inside the editor they already work in, plus the option to hand longer jobs to a cloud agent that works on a branch and opens a pull request. The cloud agent connects to GitHub Cloud and Enterprise Server, GitLab Cloud and Self-Hosted, Bitbucket Cloud and Azure DevOps, so it suits teams whose repos are not all on one host. Teams tiers add SAML/OIDC SSO, usage analytics and a team-wide privacy setting that new members inherit.
Anyone who needs a predictable monthly bill should be careful: Cursor's own pricing docs quantify nothing about included usage (only 'Limited', 'Extended' and 'Generous'), while estimating typical spend at $60-100 a month for daily agent users and $200+ for power users, against a $20 Pro price, with overflow billed in arrears at API rates. The free Hobby tier is 'Limited Agent requests' with no published number, so it is a taster rather than a working allowance. Teams with a hard no-training requirement need to check the setting themselves: Cursor's commitment is conditional, stated as 'When enabled, we will not train on your data' under Privacy Mode, and the security page publishes no retention periods or hosting regions.
Pros
Cons
Only Privacy Mode answers this. Cursor's security page says that when Privacy Mode is enabled, it will not train on your data, and that it relies on technical controls plus contractual requirements with its model providers. Privacy Mode is available on free and Pro accounts and a team or enterprise admin can turn it on for everyone, with new members inheriting the team setting. With it off, no equivalent commitment is published.
The sticker prices are Pro $20, Pro Plus $60, Ultra $200, Teams Standard $40 per user and Teams Premium $120 per user. The real figure depends on usage, and Cursor's own docs estimate $60-100 a month for daily agent users and $200 or more for people running multiple agents or automation. Usage past your plan's included amount continues at API rates, billed in arrears.
Yes. Cursor's cloud agents run in isolated cloud VMs, work on a separate branch, push to your repo and can open a pull request in each repo they change, including multi-repo setups. You can start them by commenting @cursor on a GitHub or Bitbucket pull request or issue, from Slack or Linear, or through the API. Review still sits with you before merge.
Cursor's FAQ says every plan includes a set amount of model usage, and when that runs out on-demand usage lets you keep going, billed in arrears at the same API rates. Alternatively you move to a higher tier. Cursor states requests are never downgraded in quality or speed when you are on on-demand usage. The included amount itself is not published as a number.
Facts sourced from the vendor’s own site, last verified 2026-10-10.
The Weekly Source
One email a week. New listings, repriced tools, and the ones we would skip - including this one if its pricing changes.
