Introducing the Official Figranium TypeScript and JavaScript SDK
Today, I’m introducing @figranium/sdk, the official TypeScript and JavaScript client for Figranium.
The SDK gives developers a straightforward way to bring Figranium into their own applications, services, internal tools, and automation pipelines. Instead of manually wiring together API requests, authentication, errors, and live updates, you can work through one consistent package designed around how Figranium actually operates.
This is not a return to distributing the Figranium browser runtime through npm. The engine remains self-hosted and containerized. Your Figranium instance continues to own the browser, sessions, proxies, execution environment, and automation lifecycle. The SDK is the connection between that engine and the software you are building around it.
Why the SDK Matters#
Figranium has grown beyond a visual workflow builder. It can act as an automation layer for AI agents, data pipelines, backend services, internal dashboards, and products that need reliable interaction with websites.
The API already made those integrations possible. The SDK makes them much easier to build and maintain.
With @figranium/sdk, an application can create and update tasks, start browser automations, pass in variables, follow live execution progress, inspect results, manage schedules, and work with captures or credentials without recreating the same integration layer every time.
The goal is simple: let developers focus on what their automation should accomplish instead of the repetitive plumbing around it.
Built for TypeScript and JavaScript#
The SDK is TypeScript-first, with clear types for tasks, variables, actions, schedules, results, and errors. That gives TypeScript projects useful guidance directly inside the editor and catches many integration mistakes before a workflow runs.
JavaScript users get the same client without needing TypeScript. The package supports both modern imports and CommonJS, works in Node.js 18 or newer and modern browsers, and has no runtime dependencies.
Whether you are adding Figranium to a small script or a larger product, the setup stays lightweight.
Build and Run Tasks from Your Application#
Figranium tasks can now be created in code with the same actions and variables used by the visual editor.
The SDK includes helpers for common browser interactions, content extraction, control flow, HTTP requests, screenshots, task chaining, downloads, and CAPTCHA solving. It also handles Figranium’s variable template format, so reusable tasks can receive different values on every run.
This opens up practical workflows such as:
- Creating customer-specific tasks from an application
- Triggering a saved automation from a backend service
- Passing search terms, account details, or job settings at runtime
- Scheduling and monitoring recurring extraction jobs
- Streaming live execution updates into a dashboard
- Launching a visible browser session for inspection or intervention
- Sending structured results into another database or system
The SDK does not replace the visual editor. It complements it. You can design tasks visually, call them from code, or generate them entirely inside your own application.
One Client for the Wider Figranium Platform#
The first release is designed around the full Figranium API rather than a single execution endpoint.
It includes dedicated areas for tasks, executions, schedules, captures, credentials, managed browser sessions, settings, authentication, and service health. It also supports direct scrape, agent, and headful runs when you do not need to save a task first.
For long-running work, live execution and selector events can be followed as they happen. Requests can be cancelled, time limits can be adjusted, and errors carry useful Figranium details instead of collapsing into generic failures.
Those capabilities make the SDK suitable for real application flows where users expect progress, cancellation, and meaningful recovery—not just a request that disappears into the background.
Designed to Stay Small and Predictable#
@figranium/sdk has zero runtime dependencies. It ships with support for both major JavaScript module formats, includes its TypeScript definitions, and is tested across authentication, task execution, errors, timeouts, resource routing, and live event streams.
Releases are automated through GitHub Actions with npm provenance, allowing published packages to be traced back to their source and release workflow.
The SDK is open source under the Apache 2.0 license.
Complete Documentation Is Now Live#
Alongside the SDK, I have added a complete documentation section at official Figranium SDK documentation.
It includes:
- Installation for npm, pnpm, yarn, and bun
- A guided quickstart
- Client configuration and authentication
- Task actions and runtime variables
- Request options, errors, and live streaming
- Individual references for every SDK resource
The blog post is the introduction; the docs are where the implementation details live.
Getting Started#
Install the SDK with npm install @figranium/sdk, point the client at your self-hosted Figranium instance, and provide the API key generated from Figranium’s settings.
The quickstart walks through creating, saving, and running a typed task from start to finish.
The source code and examples are available in the Figranium JavaScript SDK repository. The self-hosted automation engine remains available in the main Figranium repository.
This release gives developers a stable way to make browser automation part of their own product architecture while keeping the browser, data, and execution environment under their control.