[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-75032":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":16,"subscribersCount":16,"size":16,"stars1d":17,"stars7d":18,"stars30d":19,"stars90d":16,"forks30d":16,"starsTrendScore":20,"compositeScore":21,"rankGlobal":10,"rankLanguage":10,"license":22,"archived":23,"fork":23,"defaultBranch":24,"hasWiki":25,"hasPages":23,"topics":26,"createdAt":10,"pushedAt":10,"updatedAt":27,"readmeContent":28,"aiSummary":29,"trendingCount":16,"starSnapshotCount":16,"syncStatus":17,"lastSyncTime":30,"discoverSource":31},75032,"web-agent","firecrawl\u002Fweb-agent","firecrawl","🔥 Open-source web data agent optimized for structured web research","https:\u002F\u002Ffirecrawl.dev",null,"TypeScript",1118,152,5,1,0,2,9,36,6,19.55,"MIT License",false,"main",true,[],"2026-06-12 02:03:31","# Firecrawl Web Agent\n\n[![License: MIT](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-MIT-blue.svg)](.\u002FLICENSE)\n\n\u003Cimg src=\".internal\u002Fagent.jpg\" alt=\"Firecrawl Agent\" \u002F>\n\n\u003Cimg src=\"https:\u002F\u002Fmedia1.giphy.com\u002Fmedia\u002Fv1.Y2lkPTc5MGI3NjExcWhub2Jmd3NvejdhaTFsb3RvZWtpb2Q3cDVpN2pzYjVqeTgxdDEwbiZlcD12MV9pbnRlcm5hbF9naWZfYnlfaWQmY3Q9Zw\u002FCVyWVobjHwYGJiRz6r\u002Fgiphy.gif\" alt=\"Firecrawl Agent Demo\" width=\"100%\" \u002F>\n\nFirecrawl runs a research-grade autonomous agent at [firecrawl.dev\u002Fapp\u002Fagent](https:\u002F\u002Ffirecrawl.dev\u002Fapp\u002Fagent), powered by [Spark 1](https:\u002F\u002Fdocs.firecrawl.dev\u002Ffeatures\u002Fmodels) models optimized for structured web research. This repo gives you the open-source foundation to build your own — fork it, swap models, add skills, and deploy however you want.\n\n## Get started\n\n```bash\n# 1. Install the Firecrawl CLI and authenticate\nnpx -y firecrawl-cli@latest init -y --browser\n\n# 2. Scaffold an agent project\nfirecrawl create agent -t next\n```\n\n## Open Source\n\nEach layer builds on the one below it. Start at the top for a ready-to-use app, or go lower in the stack for finer control over the primitives.\n\n| Layer | Description | Get started |\n|:---:|---|---|\n| [**Next.js Template**](.\u002Fagent-templates\u002Fnext\u002F) | Chat UI, streaming, Skills, Subagents, structured output | `firecrawl create agent -t next` |\n| [**Express Template**](.\u002Fagent-templates\u002Fexpress\u002F) | API server with Skills, Subagents, structured output | `firecrawl create agent -t express` |\n| ↑ | | |\n| [**Agent Core**](.\u002Fagent-core\u002F) | Orchestrator built on [Deep Agents](https:\u002F\u002Fdocs.langchain.com\u002Foss\u002Fjavascript\u002Fdeepagents\u002Foverview) (LangChain). Skills, Subagents, structured output | `firecrawl create agent -t library` |\n| ↑ | | |\n| [**Firecrawl AI SDK**](https:\u002F\u002Fnpmjs.com\u002Fpackage\u002Ffirecrawl-aisdk) | Search, Scrape, Interact as Vercel AI SDK tools | `npm i firecrawl-aisdk` |\n| ↑ | | |\n| [**Firecrawl SDK**](https:\u002F\u002Fwww.npmjs.com\u002Fpackage\u002F@mendable\u002Ffirecrawl-js) | Core API client for Scrape, Search, Crawl, Extract | `npm i @mendable\u002Ffirecrawl-js` |\n| ↑ | | |\n| [**API Reference**](https:\u002F\u002Fdocs.firecrawl.dev\u002Fapi-reference\u002Fv2-introduction) | REST API, use from any language | [docs.firecrawl.dev](https:\u002F\u002Fdocs.firecrawl.dev) |\n\n### Examples\n\n| Level | Examples |\n|---|---|\n| Next.js | [Full template](.\u002Fagent-templates\u002Fnext\u002F) |\n| Express | [API server](.\u002Fagent-templates\u002Fexpress\u002F) |\n| Agent Core | [Basic](.\u002Fagent-core\u002Fexamples\u002F1-basic.ts) · [Structured output](.\u002Fagent-core\u002Fexamples\u002F2-structured-output.ts) · [Parallel Subagents](.\u002Fagent-core\u002Fexamples\u002F3-parallel-subagents.ts) · [With Skills](.\u002Fagent-core\u002Fexamples\u002F4-with-skills.ts) · [Streaming](.\u002Fagent-core\u002Fexamples\u002F5-streaming.ts) |\n| Firecrawl AI SDK | [npmjs.com\u002Fpackage\u002Ffirecrawl-aisdk](https:\u002F\u002Fnpmjs.com\u002Fpackage\u002Ffirecrawl-aisdk) |\n\n## How it works\n\nThe agent combines web tools with an AI model in a loop — it plans, acts, observes, and repeats until the task is done. The harness is [Deep Agents](https:\u002F\u002Fdocs.langchain.com\u002Foss\u002Fjavascript\u002Fdeepagents\u002Foverview) (from LangChain), which gives us the plan-act loop, parallel `task` sub-agent spawning, and on-demand SKILL.md loading out of the box. Our `agent-core` wires Firecrawl's tools into that runtime and layers on structured output and streaming.\n\n- **Harness** — [Deep Agents](https:\u002F\u002Fdocs.langchain.com\u002Foss\u002Fjavascript\u002Fdeepagents\u002Foverview). Provides the agent loop, sub-agent spawning, skills loading, and context management.\n- **Tools** — Search, Scrape, Interact (browser automation), bash. Powered by [firecrawl-aisdk](https:\u002F\u002Fwww.npmjs.com\u002Fpackage\u002Ffirecrawl-aisdk).\n- **Skills** — reusable SKILL.md playbooks. Auto-discovered from `agent-core\u002Fsrc\u002Fskills\u002Fdefinitions\u002F`, loaded on demand via Deep Agents' skills middleware.\n- **Subagents** — parallel workers for independent tasks, spawned via Deep Agents' `task` tool. Each has its own tool set and session state (e.g. an isolated interact browser session).\n- **Output** — structured results via `formatOutput` (JSON) and data processing via `bashExec`, a set of bash tools powered by [just-bash](https:\u002F\u002Fgithub.com\u002Fvercel-labs\u002Fjust-bash).\n\n## Project structure\n\n| Directory | What's inside |\n|-----------|--------------|\n| [`agent-core\u002F`](.\u002Fagent-core\u002F) | Core agent logic, orchestrator, Skills, tools |\n| [`agent-templates\u002F`](.\u002Fagent-templates\u002F) | Deployment templates - [Next.js](.\u002Fagent-templates\u002Fnext\u002F), [Express](.\u002Fagent-templates\u002Fexpress\u002F), [Library](.\u002Fagent-templates\u002Flibrary\u002F) |\n\n## License\n\nMIT\n","Firecrawl Web Agent 是一个为结构化网络研究优化的开源网页数据代理。该项目使用 TypeScript 编写，其核心功能包括通过结合网络工具与 AI 模型循环执行任务（计划、行动、观察、重复），直至完成目标。技术特点上，它基于 LangChain 的 Deep Agents 构建，支持多种技能和子代理，并能生成结构化的输出结果。适合需要进行自动化且结构化的网络数据收集与分析的应用场景，如市场调研、竞品分析等。用户可以根据需求选择从 Next.js 或 Express 模板开始快速搭建应用，或深入底层自定义更复杂的功能。","2026-06-11 03:52:02","high_star"]