What problem does it solve?
agent-browser removes the repetitive, manual work of interacting with web pages by providing a CLI-first browser automation workflow that can navigate sites, fill forms, click elements, capture screenshots, extract content, and run tests reliably.
Core Features & Use Cases
- Interactive, compact snapshots that produce short refs for elements so agents can act with minimal context and token usage.
- Full workflow support for navigation, form filling, authentication vaults, session save/load, downloads, annotated screenshots, visual diffs, and performance profiling.
- Flexible environment support including headed/headless modes, mobile simulators, proxy configuration, CDP connection, and session isolation for concurrent scrapers or E2E tests.
- Use Case: Automate logging into a protected dashboard, extract structured data, save the authenticated state for reuse, and perform visual regression comparisons between staging and production.
Quick Start
Open https://example.com, take an interactive snapshot to identify element refs, fill the login form, and save the authenticated session state to auth.json.