Configure your preview with a coding agent (MCP)

Autonoma runs a second MCP server for onboarding: connect your coding agent and it configures your PreviewKit preview for you - reads the config, sets up apps, databases and services, asks you for any secrets it needs, deploys off your main branch, and iterates until the preview is up - while you watch, read-only, in the Autonoma UI.

A coding agent plugs into Autonoma and lights up a live preview environment

This is the agent-driven alternative to filling in the PreviewKit config form by hand. Your agent already knows your repo - its build, its services, its env - so it can do the setup faster than you can describe it. You stay in control: the UI shows every step live and hands control back the moment you take over.

Start from the Autonoma UI

You connect this server from inside onboarding, because it pins itself to the app you are setting up:

In Autonoma, start onboarding your app and open the PreviewKit config step.
Click Configure with coding agent. A modal shows the install snippet for your client and a short pairing code.
Install the server in your agent (below), then tell your agent to configure your preview and give it the pairing code.

The pairing code is how the agent knows which app to configure - you never hand it a repo name, and it can only ever touch an app in an organization you belong to.

Connection details

Setting	Value
URL	`https://api.autonoma.app/v1/mcp/onboarding`
Transport	Streamable HTTP
Authentication	OAuth - your client opens a browser to sign in to Autonoma and authorize; no API key or token to paste

Connect your coding agent

claude mcp add --transport http autonoma-onboarding https://api.autonoma.app/v1/mcp/onboarding

The first time your agent uses a tool, Claude Code opens a browser to sign in and authorize.

Add the server to ~/.cursor/mcp.json (global) or .cursor/mcp.json in your project:

{
  "mcpServers": {
    "autonoma-onboarding": {
      "url": "https://api.autonoma.app/v1/mcp/onboarding"
    }
  }
}

Reload Cursor, then complete the browser sign-in when prompted from Settings → MCP.

Add the server to ~/.codeium/windsurf/mcp_config.json:

{
  "mcpServers": {
    "autonoma-onboarding": {
      "serverUrl": "https://api.autonoma.app/v1/mcp/onboarding"
    }
  }
}

Refresh MCP servers from Cascade's settings, then complete the browser sign-in.

Any MCP client that speaks Streamable HTTP can connect with the connection details above. For a client that only supports STDIO, bridge to the remote server with mcp-remote:

{
  "mcpServers": {
    "autonoma-onboarding": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://api.autonoma.app/v1/mcp/onboarding"]
    }
  }
}

How the agent configures your preview

Pair, then loop: edit config, deploy, poll status, and retry until the preview is live - while you watch read-only

Once paired, the agent works a loop, and the Autonoma UI streams every step:

Pair - pair(code) claims your app's config for the agent and returns the current config.
Read and edit - get_config / apply_config read and save the full PreviewKit config document. A save is validated; an invalid document returns the errors to fix.
Deploy - trigger_deploy deploys your main branch as the preview.
Watch and iterate - get_session_status returns the deploy status, the preview URL, and diagnostics. The agent polls it, reads the diagnostics if the preview fails, fixes the config, and deploys again - until it comes up.

While the agent holds the config, the Autonoma UI is read-only and shows what it is doing in real time. You can press Take over at any point to reclaim the config and continue by hand; the agent stands down immediately. If the agent goes idle, control returns to you automatically.

Secrets stay yours

The agent sends only key names; you enter the values in the UI and they reach Autonoma encrypted - the agent never sees them

When your app needs a secret value the agent does not have - an API key, a database URL, a token - it calls request_env with only the key names. That surfaces a prompt in the Autonoma UI where you enter the values. The values go straight from your browser to Autonoma, encrypted; the agent never sees them and there is no tool that accepts a secret value.

The agent should always ask you first whether to set env from your .env (you paste it into the UI) or to configure it yourself - so you decide, every time, what Autonoma stores.

What your agent can do

Tool	Input	Does
`pair`	code	Claims the app's config using the pairing code from the UI; returns its `applicationId` and current config
`get_config`	applicationId	Reads the current PreviewKit config document
`apply_config`	applicationId, document	Saves the full config document (validated; errors returned to fix). Never carries secret values
`request_env`	applicationId, keys, appName	Asks you to enter secret values in the UI (keys only - the agent never sees the values)
`trigger_deploy`	applicationId	Deploys the main branch as the preview
`get_session_status`	applicationId	The single polling tool: control state, any pending request, deploy status, preview URL, and diagnostics

Troubleshooting

The tools do not show up. Confirm the client connected (e.g. claude mcp list) and that you completed the browser sign-in.

The pairing code does not work. Codes are single-use and short-lived. Reopen Configure with coding agent in the UI to mint a fresh one.

The UI went read-only and I want to edit by hand. Press Take over - the agent stands down and the form becomes editable again.