Overview & schema

How .trevize/settings.json configures a repository for Trevize — where it lives, how it's read, and its top-level keys.

A repository tells Trevize how to set itself up, run, and preview through a single versioned file: .trevize/settings.json. It lives in your repo, so the configuration travels with your code and applies to every sandbox Trevize creates for a session, task, or automation.

Where it lives

Put the file at .trevize/settings.json in the root of your repository.

When a sandbox is created, Trevize reads this file for the repository and uses it to run lifecycle hooks, start preview processes, and provision resources.

Top-level keys

{
  "$schema": "https://trevize.dev/schemas/settings.json",
  "workspaceId": "your-workspace-id",
  "projectId": "your-org/your-repo",
  "hooks": {},
  "preview": {},
  "resources": {}
}

Key	Purpose
`$schema`	Points at the JSON schema so editors can offer autocomplete and validation. Optional.
`workspaceId`	Identifies the Trevize workspace this project belongs to (used by the MCP integration).
`projectId`	Your `org/repo` identifier (can be auto-detected from the git remote).
`hooks`	Commands to run at sandbox lifecycle points. See Hooks.
`preview`	Preview targets and processes that serve your app. See Preview.
`resources`	Managed resources (e.g. a database) mapped into the sandbox. See Resources.

All keys are optional — include only what your repository needs. hooks, preview, and resources are the parts that shape what happens inside a sandbox.

Hooks

Run install/setup commands when the sandbox is ready. Configure hooks.

Preview

Serve your app so reviewers get a live preview. Configure preview.

Resources

Wire a database branch into the sandbox. Configure resources.

Annotated example

A complete, walked-through settings.json. See the example.

For a field-by-field table, see the settings.json reference.