# Browserbeam > Browserbeam is a REST API that gives AI agents and automation code control over real browsers. Send JSON, get structured page data back: markdown content, stable element refs, form metadata, change diffs, and screenshots. No Playwright or Puppeteer to manage. Browserbeam replaces raw HTML scraping with structured, token-efficient browser output. Every response includes a stability signal, interactive element refs (e1, e2, e3), and a diff of what changed since the last action. The API supports navigation, clicking, form filling, JavaScript execution, AI-powered extraction, scroll collection, screenshots, and PDF generation. Base URL: `https://api.browserbeam.com/v1/` Authentication: `Authorization: Bearer ` ## Docs - [API Documentation](https://browserbeam.com/docs/): Full API reference covering sessions, steps, element targeting, response format, errors, rate limits, and workflows with code examples in cURL, Python, TypeScript, and Ruby. - [Billing & Credits Reference](https://browserbeam.com/docs/billing/): Credit rates for runtime, proxy bandwidth, AI selectors, and CAPTCHA solving. Includes the formulas Browserbeam uses to bill every session. - [OpenAPI Specification](https://browserbeam.com/openapi.yaml): Machine-readable OpenAPI 3.1 schema for the public API. ## SDKs & Integrations - [Python SDK](https://browserbeam.com/integrations/python-sdk/): Install with pip. Sync and async clients, full type hints, session lifecycle helpers. - [TypeScript SDK](https://browserbeam.com/integrations/typescript-sdk/): Install with npm. Full TypeScript types, ESM and CJS builds. - [Ruby SDK](https://browserbeam.com/integrations/ruby-sdk/): Install with gem. Block-based sessions, Struct-based response types. - [MCP Server](https://browserbeam.com/integrations/mcp-server/): Use Browserbeam as tools in Cursor, Claude Desktop, and Windsurf via the Model Context Protocol. ## Guides - [Getting Started with Browserbeam's Python SDK](https://browserbeam.com/blog/python-sdk-getting-started/): Install the SDK, create sessions, navigate pages, fill forms, extract data, and take screenshots with working Python code. - [OpenAI Agents SDK + Browserbeam](https://browserbeam.com/blog/openai-agents-sdk-browserbeam/): Build OpenAI Agents SDK tools that browse, click, and extract structured data from any web page. - [Browserbeam & LangChain Integration](https://browserbeam.com/blog/browserbeam-langchain-integration/): Wrap Browserbeam methods as LangChain tools and wire them into a LangGraph ReAct agent. - [MCP Meets Browserbeam](https://browserbeam.com/blog/mcp-meets-browserbeam/): Set up the Browserbeam MCP server, connect it to your IDE, and run browser tasks from your AI coding assistant. - [Claude Code + Browserbeam MCP](https://browserbeam.com/blog/claude-code-browserbeam-mcp/): Configure the Browserbeam MCP server for Claude Code with structured browser output and declarative extraction. - [Structured Web Scraping Guide](https://browserbeam.com/blog/structured-web-scraping-guide/): Define declarative schemas and extract typed JSON from any page, including JavaScript-rendered SPAs. - [LLM-Powered Browser Automation Tutorial](https://browserbeam.com/blog/llm-powered-browser-automation/): Build an LLM agent loop that navigates, observes, and acts on web pages using function calling. - [Building Intelligent Web Agents](https://browserbeam.com/blog/building-intelligent-web-agents/): Architecture patterns for AI browser agents: ReAct loops, multi-step workflows, and error recovery. ## Use Cases - [Web Scraping with Python](https://browserbeam.com/blog/web-scraping-python/): Step-by-step Python scraping with BeautifulSoup, Selenium, and Browserbeam using real websites. - [Web Scraping with Node.js](https://browserbeam.com/blog/web-scraping-nodejs/): Node.js scraping with Cheerio, Puppeteer, Playwright, and Browserbeam with working code examples. - [Building a Web Scraping Agent in Python](https://browserbeam.com/blog/web-scraping-agent-python/): Build a Python agent that logs in, navigates multi-page flows, and extracts structured data. - [How to Build a Price Monitoring Bot](https://browserbeam.com/blog/price-monitoring-bot/): Scrape prices, store history in SQLite, detect changes, and send Slack alerts. - [Build a Competitive Intelligence Agent](https://browserbeam.com/blog/competitive-intelligence-agent/): Automate competitor monitoring with browser sessions and GPT-powered analysis. - [How to Build a Web Crawler](https://browserbeam.com/blog/build-web-crawler/): Crawler architecture, robots.txt handling, database storage, and production patterns in Python, JavaScript, and Ruby. - [Data Extraction: From Browser to JSON](https://browserbeam.com/blog/data-extraction-browser-to-json/): Extract engine internals, schema design, nested data handling, and performance benchmarks. ## Concepts - [Raw HTML vs Structured Output](https://browserbeam.com/blog/raw-html-vs-structured-output/): Why AI agents prefer structured browser output over raw HTML, with token cost comparisons. - [Browser as a Service](https://browserbeam.com/blog/browser-as-a-service/): Managed vs self-hosted browser infrastructure and how structured output changes the economics. - [Scaling Web Automation](https://browserbeam.com/blog/scaling-web-automation/): Queue architecture, capacity planning, cost optimization, and monitoring for 100 to 100,000 sessions. - [Securing AI Browser Agents](https://browserbeam.com/blog/securing-ai-browser-agents/): Session isolation, credential management, prompt injection defense, and GDPR compliance. ## Optional - [How to Scrape Amazon](https://browserbeam.com/blog/scrape-amazon/): Extract prices, reviews, and product data with residential proxies. - [Scrape eBay](https://browserbeam.com/blog/scrape-ebay/): Extract prices, conditions, and seller ratings with CSS selectors or markdown. - [How to Scrape Reddit](https://browserbeam.com/blog/scrape-reddit/): Extract posts, comments, and author data with residential proxy bypass. - [Scrape Airbnb Listings](https://browserbeam.com/blog/scrape-airbnb/): Extract nightly rates, host ratings, and amenities. - [Scrape Shopify Stores](https://browserbeam.com/blog/scrape-shopify/): Use the hidden /products.json endpoint and the browser API for what JSON misses. - [Extract Wikipedia Data](https://browserbeam.com/blog/scrape-wikipedia/): Pull articles, infoboxes, and references with five lines of code. - [How to Scrape YouTube](https://browserbeam.com/blog/scrape-youtube/): Extract video data, channel listings, and transcripts. - [How to Scrape IMDb](https://browserbeam.com/blog/scrape-imdb/): Extract ratings, cast, and metadata via LD+JSON parsing. - [Scrape Redfin](https://browserbeam.com/blog/scrape-redfin/): Extract property prices, price history, and neighborhood stats. - [AliExpress Scraper](https://browserbeam.com/blog/scrape-aliexpress/): Extract product data, pricing tiers, and supplier info. - [Puppeteer vs Playwright vs Browserbeam](https://browserbeam.com/blog/puppeteer-vs-playwright-vs-browserbeam/): Side-by-side code, feature tables, and a decision framework. - [Top 5 Cloud Browser APIs Compared](https://browserbeam.com/blog/cloud-browser-apis-compared/): Browserbeam, Browserbase, Steel, Browserless, and Firecrawl compared on pricing and features. - [GraphQL vs REST for Agent Workflows](https://browserbeam.com/blog/graphql-vs-rest-agent-workflows/): Which API style works better for browser automation agents. - [Residential vs Datacenter Proxies](https://browserbeam.com/blog/residential-vs-datacenter-proxies/): Compare detection rates, real prices, and choose the right proxy type. - [HTML to PDF](https://browserbeam.com/blog/html-to-pdf/): Generate PDFs from any web page with working code in Python, TypeScript, Ruby, and cURL. - [Web Scraping in 2026](https://browserbeam.com/blog/web-scraping-2026/): Tools, techniques, legal guidance, and best practices for the current landscape. - [The Future of AI Agents](https://browserbeam.com/blog/future-ai-agents-browser-primitives/): Browser APIs as the new primitives for agent infrastructure.