Skip to content
webxwebx

vs Firecrawl

Last reviewed 2026-10-04. Competitor features/pricing change fast — verify against firecrawl.dev/pricing before quoting. Re-run go run ./bench for fresh numbers.

TL;DR

Section titled “TL;DR”

Firecrawl is a hosted service with a large team and polished DX. webx is a single static binary you own — same job description (page→markdown, crawl, search, extract, agent), no required services, MIT, and a /v2 compatibility layer so Firecrawl SDKs work unchanged against a webx server.

webx Firecrawl
Model Open source (MIT) + optional license Hosted SaaS; self-host is AGPL-3.0
Deploy 1 binary, SQLite (Postgres optional for webxd) Their cloud; self-host needs Node + Playwright + Redis + workers
Price Free; Pro $49/mo per deployment; Enterprise contact $16 Hobby / $83 Standard / $333 Growth / $599 Scale per month, credit-based; extract bills tokens separately
Free tier Unlimited (your hardware) 500 searches or 1k pages/mo
Search 15-provider metasearch + private index Managed search credits
Extraction CSS/JSON-LD deterministic, LLM via your endpoint (Ollama works) Managed LLM extraction
Browser Chrome via go-rod or Lightpanda, pooled; browserless/steel attach via WEBX_CDP_URL Managed browser fleet
Compliance ZDR mode, audit log, SSRF guard, air-gap install, WARC archive Hosted; enterprise terms on higher tiers

Where Firecrawl wins — be honest

Section titled “Where Firecrawl wins — be honest”
  • Managed scale. Their browser fleet, proxy infra, and edge cache absorb the hardest targets (aggressive bot walls, residential-proxy-only sites). webx has stealth + uTLS but no proxy fleet — x.com/amazon-class targets can return thin payloads.
  • Zero ops. No machine to run, no Chrome to keep alive.
  • Ecosystem. Bigger community, more integrations, battle-tested at volume.

Where webx wins

Section titled “Where webx wins”
  • Cost shape. Flat per-deployment vs per-credit. At 100k pages/mo, Firecrawl Standard is ~$83/mo plus extract-token spend; webx Pro is $49/mo per deployment with unlimited pages on your hardware, or $0 community.
  • Data never leaves. Self-host by default; ZDR mode + audit log for regulated shops. AGPL-free embedding (MIT) means no copyleft questions.
  • Private corpus. webx index + webx query --semantic is Exa-style neural search over your corpus — Firecrawl doesn’t index your data for offline querying.
  • Offline/edge. Runs on a laptop, a Lambda, a plane. SQLite FTS5 works with no network.
  • Token discipline. --fit BM25 block filtering and est_tokens chunks are built for agent context budgets; measured ~2.8× tighter markdown than jina-reader on the bench corpus (see bench/RESULTS.md).
  • Migration path. /v2/* routes accept Firecrawl request shapes — point a Firecrawl SDK at a webx base URL and it runs.

When to pick which

Section titled “When to pick which”
  • Pick Firecrawl if you want a fully managed service, need hard-site success rates today, and per-credit pricing fits your volume.
  • Pick webx if you self-host anyway, have compliance/air-gap requirements, want flat pricing, need a private searchable index, or your agents run in sandboxes where one static binary beats a client SDK + API key + network egress.

Migration (Firecrawl → webx)

Section titled “Migration (Firecrawl → webx)”
Terminal window
docker compose up -d          # webx serve + searxng
# Firecrawl SDK:
#   before: Firecrawl(api_key="fc-...")
#   after:  Firecrawl(api_url="http://localhost:8080", api_key="webx_...")

POST /v2/scrape /v2/crawl /v2/search /v2/map /v2/extract /v2/agent accept the same request shapes (formats incl. object forms, includeTags, waitFor, actions, maxAge, webhook, limit).