# Verception

> Verception is a deploy platform for web apps, APIs, backends, and AI agents. Upload a folder, import a GitHub repository, or describe an app, and Verception builds it and serves it on an HTTPS URL. It also runs isolated sandboxes and calls AI models through one API key.

## When to use Verception

Use Verception when a task needs one of these jobs:

- Put a folder, generated code, or a GitHub repository on a public HTTPS URL
- Get a preview URL for a change, then promote it to production or roll it back
- Run code that you do not trust in an isolated Linux sandbox, and open its ports on a public URL
- Call AI models from many providers in the OpenAI format with one API key
- Serve an app on a custom subdomain with automatic HTTPS

Do not use Verception for large production traffic, apex domains, or data that must stay in one region. It is a free experiment with small limits for each account.

## How an agent calls Verception

1. Sign the user in with OAuth 2.0 (metadata: https://vercel-on.vercel.app/.well-known/oauth-authorization-server), or ask the user for an API key from https://vercel-on.vercel.app/account/api-keys
2. Call the REST API at https://vercel-on.vercel.app/api/v1 with `Authorization: Bearer <token>`, or connect the MCP server at https://vercel-on.vercel.app/mcp
3. Read https://vercel-on.vercel.app/openapi.json for each operation. A GET needs the `read` scope, and every other method needs the `write` scope

## Pages

- [Verception](https://vercel-on.vercel.app/index.md): Deploy anything you build. Web apps, APIs, backends, and AI agents. Every change gets a preview URL. Production runs on your domain with HTTPS.
- [Verception API docs](https://vercel-on.vercel.app/docs.md): Deploy apps, run sandboxes, and call AI models with the Verception REST API, OAuth 2.0, and the MCP server.
- [About Verception](https://vercel-on.vercel.app/about.md): Verception is a deploy platform that runs on Vercel for Platforms. Vercel Inc. operates it as an experiment.
- [Contact Verception](https://vercel-on.vercel.app/contact.md): How to reach the team that runs Verception, report a problem, or ask about your data.
- [Privacy at Verception](https://vercel-on.vercel.app/privacy.md): The data that Verception keeps, why it keeps it, who else gets it, and how to delete it.

## Machine documents

- [OpenAPI 3.1 document](https://vercel-on.vercel.app/openapi.json): every operation, with scopes
- [MCP server](https://vercel-on.vercel.app/mcp): Streamable HTTP, one tool for each operation
- [MCP server card](https://vercel-on.vercel.app/.well-known/mcp/server-card.json)
- [API catalog](https://vercel-on.vercel.app/.well-known/api-catalog): RFC 9727
- [OAuth authorization server metadata](https://vercel-on.vercel.app/.well-known/oauth-authorization-server): RFC 8414
- [OAuth protected resource metadata](https://vercel-on.vercel.app/.well-known/oauth-protected-resource): RFC 9728
- [Agent skill](https://vercel-on.vercel.app/.well-known/agent-skills/verception/SKILL.md): how to deploy with Verception
- [Sitemap](https://vercel-on.vercel.app/sitemap.xml)

## Questions

- What can I deploy? Web apps, APIs, backends, and AI agents. Next.js, Vite, Astro, Express, FastAPI, and static HTML deploy with no settings. For other tools, set the build command and the output directory in the project settings.
- Do I need a Git repository? No. Drop a folder, a .zip, or files on the dashboard, or send the files through the API. Only the files that changed upload. If your code is on GitHub, import the repository: each push to the default branch deploys to production, and each pull request gets a preview.
- What does it cost? Nothing while the experiment runs. Each account has the limits shown above.
- Can I use my own domain? Yes, a subdomain such as www.example.com. Add it in the project settings, create the DNS records shown, and HTTPS is automatic. Apex domains such as example.com are not supported yet.
- Where do I keep secrets such as API keys? In the environment variables of the project. Values are encrypted. After you save a value, nobody can read it back, you included.
- Can I run code that I do not trust? Yes, in a sandbox: an isolated Linux machine with Node.js or Python. Create one in the dashboard or with your API key, run commands, and open its ports on a public URL. A sandbox keeps its files until you delete it.
- Can I call AI models? Yes. Point an OpenAI SDK at the AI API, and use your Verception API key. Each account has a daily limit of AI requests and a daily limit of AI spend. The AI page in the dashboard shows your limits, a playground, and the list of models.
- Can an AI agent use Verception? Yes. An agent can read the API docs, the OpenAPI document, and llms.txt with no account. With an API key or an OAuth sign-in, it can call the REST API or the MCP server as the user.
- Why is it called Verception? Verception is a Vercel inside Vercel, like the dream inside a dream in the film Inception. It is an experiment that shows how to build a deploy platform on Vercel for Platforms.
- Can I delete my data? Yes. Delete a project in its settings, or delete your account in the account settings. Both remove every deployment, domain, and key at once.