
firecrawl / web-agent
✓ Official★ 1,100A skill package that teaches your agent 6 capabilities — every one documented and browsable below, no GitHub required · by firecrawl.
Each skill below is one capability this package teaches your agent. Install the whole package, or open a skill to install just that one.
Structured side-by-side comparison of competing products. Designed for search + scrape; no interact needed for typical marketing/pricing pages.
1 file — installable on its own
Multi-source research with source triangulation and fact-checking. Use for any research task requiring 3+ sources.
1 file — installable on its own
Navigate e-commerce sites to extract products, pricing, categories, and inventory. Handles pagination, variants, and JS-heavy storefronts.
1 file — installable on its own
Extract official financials from SEC filings and cross-reference with analyst consensus. Public companies only.
4 files — installable on its own
Extract pricing tiers from a vendor's pricing page and normalize them into a consistent shape. Optimized for SaaS, API, cloud, LLM, and CDN pricing pages — all of which share similar structure but inconsistent naming.
1 file — installable on its own
Extract structured data matching a JSON schema from websites. Handles complex nested schemas, arrays, pagination, and validation. Always outputs via…
1 file — installable on its own
Firecrawl Web Agent
<img src=".internal/agent.jpg" alt="Firecrawl Agent" /> <img src="https://media1.giphy.com/media/v1.Y2lkPTc5MGI3NjExcWhub2Jmd3NvejdhaTFsb3RvZWtpb2Q3cDVpN2pzYjVqeTgxdDEwbiZlcD12MV9pbnRlcm5hbF9naWZfYnlfaWQmY3Q9Zw/CVyWVobjHwYGJiRz6r/giphy.gif" alt="Firecrawl Agent Demo" width="100%" />Firecrawl runs a research-grade autonomous agent at firecrawl.dev/app/agent, powered by Spark 1 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.
Get started
# 1. Install the Firecrawl CLI and authenticate
npx -y firecrawl-cli@latest init -y --browser
# 2. Scaffold an agent project
firecrawl create agent -t nextOpen Source
Each 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.
| Layer | Description | Get started |
|---|---|---|
| Next.js Template | Chat UI, streaming, Skills, Subagents, structured output | firecrawl create agent -t next |
| Express Template | API server with Skills, Subagents, structured output | firecrawl create agent -t express |
| ↑ | ||
| Agent Core | Orchestrator built on Deep Agents (LangChain). Skills, Subagents, structured output | firecrawl create agent -t library |
| ↑ | ||
| Firecrawl AI SDK | Search, Scrape, Interact as Vercel AI SDK tools | npm i firecrawl-aisdk |
| ↑ | ||
| Firecrawl SDK | Core API client for Scrape, Search, Crawl, Extract | npm i @mendable/firecrawl-js |
| ↑ | ||
| API Reference | REST API, use from any language | docs.firecrawl.dev |
Examples
| Level | Examples |
|---|---|
| Next.js | Full template |
| Express | API server |
| Agent Core | Basic · Structured output · Parallel Subagents · With Skills · Streaming |
| Firecrawl AI SDK | npmjs.com/package/firecrawl-aisdk |
How it works
The 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 (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.
- Harness — Deep Agents. Provides the agent loop, sub-agent spawning, skills loading, and context management.
- Tools — Search, Scrape, Interact (browser automation), bash. Powered by firecrawl-aisdk.
- Skills — reusable SKILL.md playbooks. Auto-discovered from
agent-core/src/skills/definitions/, loaded on demand via Deep Agents' skills middleware. - Subagents — parallel workers for independent tasks, spawned via Deep Agents'
tasktool. Each has its own tool set and session state (e.g. an isolated interact browser session). - Output — structured results via
formatOutput(JSON) and data processing viabashExec, a set of bash tools powered by just-bash.
Project structure
| Directory | What's inside |
|---|---|
agent-core/ | Core agent logic, orchestrator, Skills, tools |
agent-templates/ | Deployment templates - Next.js, Express, Library |
License
MIT
Install the whole package (6 skills):
npx skills add https://github.com/firecrawl/web-agentOr install a single skill:
npx skills add https://github.com/firecrawl/web-agent --skill <name>Pick the skill name from the Skills tab — each entry there installs independently.