एजेंट और मल्टी-एजेंट

Firecrawl

firecrawl/firecrawl

An open-source web context API to search, scrape, and interact with the web at scale, converting web data into clean Markdown or structured JSON for AI agents and LLMs.

आधिकारिक रिपॉज़िटरी खोलें ↗ प्रोजेक्ट होमपेज ↗
★ 161.1Kसितारे
⑂ 9.1Kफ़ोर्क्स
476खुले मुद्दे
TypeScriptभाषा
AGPL-3.0लाइसेंस
Q98संपादकीय स्कोर

प्रोजेक्ट के स्क्रीनशॉट

Firecrawl का स्क्रीनशॉट Firecrawl का स्क्रीनशॉट Firecrawl का स्क्रीनशॉट

अवलोकन

Firecrawl provides a suite of APIs and SDKs designed to extract, structure, and interact with web data. It handles complex web scraping challenges such as JavaScript-rendered pages, proxy rotation, and rate limits. The platform offers core endpoints for searching, scraping, and interacting with pages, alongside advanced features like an autonomous AI agent for data gathering, website crawling, URL mapping, and batch scraping. It outputs LLM-ready formats like Markdown and structured JSON to facilitate integration with AI applications.

प्रमुख विशेषताएँ

  • Web Search with full page content extraction
  • Single page Scrape to Markdown, HTML, JSON, or screenshots
  • Page Interact for clicking, scrolling, and writing via AI prompts
  • Autonomous Agent for automated data gathering using natural language
  • Website Crawl for scraping all URLs asynchronously
  • URL Map for instant site discovery
  • Batch Scrape for processing thousands of URLs
  • LLM-ready output formats to reduce token usage
  • Built-in handling of proxies, rate limits, and JS-blocked content
  • Media parsing for web-hosted PDFs and DOCX files

आवश्यकताएँ, इंस्टॉलेशन और त्वरित शुरुआत

Python: pip install firecrawl-py
Node.js: npm install firecrawl
Go: go get github.com/firecrawl/firecrawl/apps/go-sdk
Ruby: gem install firecrawl-sdk
.NET: dotnet add package firecrawl-sdk
PHP: composer require firecrawl/firecrawl-sdk
Rust: Add firecrawl = "2" to Cargo.toml dependencies
Elixir: Add {:firecrawl, "~> 1.0" to deps
Java: Add 'com.github.firecrawl:firecrawl-java-sdk:2.0' via Gradle/Maven with JitPack

उपयोग

Initialize the client with an API key, then call methods like search, scrape, interact, crawl, map, batch_scrape, or agent. For the Agent endpoint, provide a prompt describing the data needed and optionally a Pydantic schema for structured output. To connect to AI agents or MCP clients, use the CLI skill command (npx -y firecrawl-cli@latest init --all --browser) or configure the MCP server JSON with npx -y firecrawl-mcp.

मॉडल संगतता और उपयोग के मामले

The Agent endpoint supports model selection between 'spark-1-mini' (default, lower cost) and 'spark-1-pro' (standard cost, for complex research and critical data gathering).

लाइसेंस और जोखिम संबंधी टिप्पणियाँ

The project is primarily licensed under the GNU Affero General Public License v3.0 (AGPL-3.0). The SDKs and some UI components are licensed under the MIT License.

Editorial verification 2026-08-09: repository URL, owner, description, license and repository statistics were reviewed. License metadata: AGPL-3.0. README was fetched for the channel draft; re-check repository dependencies, releases and model terms before production use.

रिलीज़ और रखरखाव

Not stated in the repository metadata

LangChain

langchain-ai/langchain

★ 143.6KPython

RAGFlow

infiniflow/ragflow

★ 86.7KGo