Skip to main content

Developers

Control Browlix from code and AI agents

API v1, the MCP server, the CLI and the TypeScript SDK share one operation catalog with the dashboard. API keys are scoped; every write returns a receipt so retries are safe.

Open the docsDownload

  • API v1

    REST with OpenAPI 3.1, authenticated with scoped API keys.

  • MCP server

    For Claude Code, Codex, Cursor, VS Code and other AI agents; same permissions as the API key.

  • browlix CLI

    A single binary for Windows and Linux, verified with SHA-256 on install.

  • TypeScript SDK

    @browlix/sdk with types generated from OpenAPI and Playwright attach support.

Try it in a minute

Create an API key in your workspace under Settings → API keys, put it in the BROWLIX_API_KEY environment variable and call:

API v1
https://browlix.com/api/public/v1
OpenAPI
https://browlix.com/api/public/v1/openapi.json
MCP
https://browlix.com/api/public/mcp
List profiles
curl https://browlix.com/api/public/v1/profiles \
  -H "Authorization: Bearer $BROWLIX_API_KEY"

Where your code runs

Remotely

A backend, CLI or AI agent anywhere: manage profiles, proxies, sessions, flows, AI tasks, webhooks and feedback through the API or MCP.

On the worker's machine

Code on a machine with a worker attaches Playwright or Puppeteer to the browser through the local broker using a short-lived ticket. Tickets do not work from another machine.

Feedback built for AI agents

The feedback_submit tool lets an agent report a bug, friction or unclear docs after using a tool — only when there is something worth reporting. Tickets are scrubbed of secrets, deduplicated, and the Browlix team replies right in the ticket.