Cursor can now publish web apps with Postgres and a custom domain on its own

Cursor's cloud agent can now take a project all the way to publication: it creates the structure, spins up Postgres, connects authentication through Cursor SSO, and provides a link to the application on a custom domain. Before deployment, it checks the code for vulnerabilities and exposed secrets. It works with both new and existing GitHub repositories.

AI-processed from TG: @ai_machinelearning_big_data (Machinelearning); edited by Hamidun News
Cursor can now publish web apps with Postgres and a custom domain on its own
Source: TG: @ai_machinelearning_big_data (Machinelearning). Collage: Hamidun News.
Listen to article
Article Fact Sheet & Key Figures
ParameterDetails
Primary sourceTG: @ai_machinelearning_big_data (Machinelearning) ↗
Editor-in-Chief
Key metric #1The agent sets up Postgres, SSO, and hosting itself
Key metric #2It checks for vulnerabilities and secrets before publication
Key metric #3A paid plan with cloud agents is required

It is a familiar pain point: the code is written, but turning it into a live application still takes half a day of hassle. You need to find hosting, spin up a database, configure authentication, connect a domain, and remember to hide your keys. Cursor has set out to close this gap by handing the entire process to its cloud agent.

What the agent can do

Cursor's cloud agent can now create the project structure, spin up a Postgres database, configure authentication through Cursor SSO, and provide hosting on its own. At the end of the process, the user receives a link to a working application with a custom domain.

Users can publish both new projects and existing GitHub and Cursor Origin repositories. GitLab support is promised for later. If your code lives there, you will have to wait.

Pre-deployment checks

The most practical part is not the hosting but the safety gate. Before publication, the code is checked for vulnerabilities and exposed secrets, including the very API keys developers routinely commit by accident. Until the checks pass, the agent explains exactly what needs to be fixed and blocks publication.

Imagine an internal tool for a sales department: a form, a database table, and sign-in through a corporate account. Previously, this was a weekend mini-project. Now it is a task you can describe to the agent, receive a link, and then inspect the result yourself.

Limitations and takeaways

The feature is available only on paid plans that include cloud agents. Authentication is tied to Cursor SSO, while hosting is provided by Cursor itself. This is convenient for prototypes and internal tools, but production systems with strict infrastructure, compliance, and portability requirements will need to weigh the risk of vendor lock-in.

What should you do in practice? Developers should try the agent on a small internal service and see what issues it flags during its secrets scan. Product teams can use it for rapid prototypes that can be shown to users today rather than a week from now. Founders should note that the boundary between a code editor and a hosting platform is disappearing, which means Cursor is now competing not only with other editors but also with deployment services.

The main open question is how reliably the agent configures the database and access permissions without supervision. For now, it is sensible to treat the output as a draft that requires review, especially if the application contains real customer data.

Story

Key Questions Answered

What the agent can do?

Cursor's cloud agent can now create the project structure, spin up a Postgres database, configure authentication through Cursor SSO, and provide hosting on its own. At the end of the process, the user receives a link to a working application with a custom domain. Users can publish both new projects and existing GitHub and Cursor Origin repositories. GitLab support is promised for later. If your code lives there, you will have to wait.

What are the key details on Pre-deployment checks?

The most practical part is not the hosting but the safety gate. Before publication, the code is checked for vulnerabilities and exposed secrets, including the very API keys developers routinely commit by accident. Until the checks pass, the agent explains exactly what needs to be fixed and blocks publication. Imagine an internal tool for a sales department: a form, a database table, and sign-in through a corporate account. Previously, this was a weekend mini-project.

What are the key details on Limitations and takeaways?

The feature is available only on paid plans that include cloud agents. Authentication is tied to Cursor SSO, while hosting is provided by Cursor itself. This is convenient for prototypes and internal tools, but production systems with strict infrastructure, compliance, and portability requirements will need to weigh the risk of vendor lock-in. What should you do in practice? Developers should try the agent on a small internal service and see what issues it flags during its secrets scan.

ZK
Zhemal Khamidun — CPO at Alpina Digital, product manager of AlpinaGPT. Building AI products since 2019. Hamidun News is a pet project that grew into a production aggregator with 6,000+ readers.
✦ Hands-On AI Engineering · Hamidun Academy

Build & Ship Production AI Agents Without Boilerplate Code

Start your free 72-hour trial at Hamidun Academy (academy.hamidun.com) — interactive lessons on Claude Code & Cursor, live Tech Radar, and production agent templates.

Start 72h Free Trial →Vibe-Coding Intensive
What do you think?
Loading comments…