webrobot-sdk-nodejs

Install the webrobot-sdk package and create a WebRobotClient with your baseUrl and either an API key or JWT. The client exposes curated methods like projectsList, jobExecute, and jobGetLogs for the most common operations.

Until npm publication, install directly from GitHub with npm install git+https://github.com/WebRobot-Ltd/webrobot-sdk.git#main:nodejs-sdk. Once published, use npm install webrobot-sdk. Node 18 or later is required since it uses native fetch.

Use the Node.js SDK for Express, Fastify, Next.js, or browser and Electron apps. Use the webrobot CLI for shell scripts and CI pipelines. Use MCP tools when working inline inside Claude Code or Cursor.

The curated WebRobotClient covers about 40 of the most-used endpoints. For the remaining 100+ endpoints, including admin, billing, and partner-vertical paths, use a generic fetch-based apiCall helper with path templating.

Yes, the client is fetch-based with no Node-specific dependencies, so it runs in browsers and Electron. However, never embed API keys in browser bundles; route calls through a server-side proxy or use short-lived JWTs.

Catch WebRobotError from the curated client to access the HTTP status code and raw response body. For generic fetch-based calls, the thrown Error includes the status code in its message, which you can wrap in your own error type.