Reasonkit Web
High-performance MCP server for browser automation, web capture, and content extraction. Rust-powered CDP client for AI agents.
Ask AI about Reasonkit Web
Powered by Claude · Grounded in docs
I know everything about Reasonkit Web. Ask me about installation, configuration, usage, or troubleshooting.
0/500
Reviews
Documentation
ReasonKit Web (Rust Edition)
High-Performance Web Sensing & Browser Automation Layer Rust-Native Implementation
The Eyes and Ears of AI Reasoning - Now Blazingly Fast
Note: This is the Rust implementation of the ReasonKit Web Sensing layer. It supersedes the legacy Python prototype for performance-critical deployments.
Web sensing and browser automation layer for ReasonKit. Implements the Model Context Protocol (MCP) for seamless web interactions with AI reasoning systems, powered by Rust and ChromiumOxide.
Features
Use Cases
Performance
Installation
Universal Installer (Recommended)
Installs all 4 ReasonKit projects together:
curl -fsSL https://get.reasonkit.sh | bash -s -- --with-web
Platform & Shell Support:
- ✅ All platforms (Linux/macOS/Windows/WSL)
- ✅ All shells (Bash/Zsh/Fish/Nu/PowerShell/Elvish)
- ✅ Auto-detects shell and configures PATH
- ✅ Beautiful progress visualization
Cargo (Rust Library)
Add to your Cargo.toml:
[dependencies]
reasonkit-web = "0.1"
tokio = { version = "1", features = ["full"] }
Quick Start
As a Library
use reasonkit_web::BrowserController;
#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
// Create browser controller
let controller = BrowserController::new().await?;
// Navigate to a page
let page = controller.navigate("https://example.com").await?;
// Extract content
let content = page.content().await?;
println!("Page content: {}", content);
Ok(())
}
As an MCP Server
# Build the MCP server binary
cargo build --release
# Run the MCP server
./target/release/reasonkit-web
# Or with cargo
cargo run --release
Content Extraction
use reasonkit_web::{ContentExtractor, MetadataExtractor};
// Extract structured content from HTML
let html = "<html><body><h1>Title</h1><p>Content</p></body></html>";
let extractor = ContentExtractor::new();
let content = extractor.extract(html)?;
// Extract metadata
let meta_extractor = MetadataExtractor::new();
let metadata = meta_extractor.extract(html)?;
API Workflow
Architecture
The ReasonKit Web layer implements a high-performance web sensing architecture designed for AI reasoning systems:
System Topology
Core Architecture
The Rust Engine
Key Design Principles
Performance-First: Built in Rust with async/await for maximum throughput Protocol-Driven: Implements Model Context Protocol (MCP) for AI integration Modular Design: Separates capture, extraction, and memory for flexibility Security-Focused: Headless browser isolation and content sanitization
Integration Flow
Security & Privacy
ReasonKit Web is built with a "GDPR by Default" philosophy, ensuring that AI reasoning systems can interact with the web safely and compliantly.
Key Security Features
- PII Redaction Engine: Automatically detects and redacts sensitive information (emails, phone numbers, secrets) before it leaves the secure runtime.
- Headless Isolation: Browser sessions run in isolated containers with no persistent state.
- Content Sanitization: Malicious scripts and trackers are stripped at the DOM level.
For detailed technical specifications, see docs/ARCHITECTURE.md.
Technology Stack
Version & Maturity
| Component | Status | Notes |
|---|---|---|
| MCP Server | ✅ Stable | Model Context Protocol integration |
| CDP Client | ✅ Stable | Chrome DevTools Protocol via chromiumoxide |
| Content Extraction | ✅ Stable | HTML parsing, metadata extraction |
| PII Redaction | ✅ Stable | GDPR-compliant data handling |
| WASM Support | 🔶 Beta | Browser-native execution |
Current Version: v0.1.3 | CHANGELOG | Releases
Verify Installation
# Build and run MCP server
cargo build --release
./target/release/reasonkit-web --help
# Or test as a library
cargo test
License
Apache License 2.0 - see LICENSE
