epiral/bb-browser
Bb Browser
bb-browser is an authenticated web scraper and browser automation CLI that also functions as an MCP server for AI coding tools. It treats the browser as a programmable runtime environment, enabling AI agents to control a live Chrome instance through a standard protocol while leveraging existing login sessions for authenticated actions.
The project distinguishes itself through a dual CLI and MCP interface, allowing both direct command-line control and AI-driven browser manipulation. It includes a parallel multi-platform query engine that executes simultaneous searches across multiple websites, aggregating structured JSON results. A plugin-based site adapter system supports over 100 pre-configured commands across 36 platforms, with the ability to create custom adapters by reverse-engineering website APIs.
The tool provides browser automation capabilities for executing actions like opening URLs, clicking elements, filling inputs, and capturing network traffic. Output is delivered as structured JSON that can be filtered and transformed inline using jq expressions. The system supports cross-platform research by querying multiple platforms on a single topic in parallel, returning combined results for broader information gathering.
Features
- Authenticated Web Data Extraction - Executes pre-built commands across 36 platforms to search and extract data using the browser's authenticated session.
- Agentic Browser Controllers - Enables AI coding tools to control a browser through an MCP server for autonomous navigation and interaction.
- Browser Control Servers - Provides an MCP server that lets AI coding tools control a browser through a standard protocol.
- MCP Servers - Ships an MCP server implementation that lets AI coding tools control a browser through the Model Context Protocol.
- Multi-Website Query Aggregators - Queries multiple websites simultaneously on a single topic, returning combined structured results for broader information gathering.
- Browser Automation - Executes browser automation tasks like opening URLs, clicking elements, filling inputs, and capturing network traffic.
- Browser Automation CLIs - Controls a live Chrome browser from the command line for automated web interactions using existing login sessions.
- CLI-Controllable Browsers - Controls a live Chrome browser from the command line using existing login sessions for authenticated interactions.
- CLI and MCP Interfaces - Provides both a command-line interface and an MCP server for the same browser automation functionality.
- Pre-Configured Site Commands - Provides over 100 pre-configured commands across 36 platforms for searching and extracting data from websites.
- Browser Session Token Reuse - Leverages existing browser login sessions to perform authenticated actions without re-entering credentials.
- Site Adapter Frameworks - Supports a plugin-based system for creating custom website adapters by reverse-engineering APIs.
- Cross-Platform Query Aggregators - Executes simultaneous queries across multiple websites, aggregating results for efficient cross-platform research.
- Browser Automation - Executes browser actions such as opening URLs, clicking elements, filling inputs, and capturing network traffic.
- Browser Automation APIs - Exposes browser automation as a programmable API layer, treating the browser as a runtime environment for executing commands.
- Chrome DevTools Protocols - Controls a live Chrome instance via the Chrome DevTools Protocol for direct DOM, network, and JavaScript manipulation.
- Multi-Platform Query Engines - Queries multiple platforms on a single topic in parallel, returning combined structured JSON results.
- jq-Transformable Outputs - Outputs structured JSON results that can be filtered and transformed inline using jq expressions.
- Website API Adapters - Creates new site adapters by reverse-engineering a website's API, with AI agents autonomously writing and submitting adapters.
- Developer Tools - Tool enabling agents to control real Chrome browser sessions.
