/ scrape · 124 domains tracked

Your request finds the cheapest path.

DreamScrape routes every request through the tier that actually works — plain HTTP, JA4-fingerprinted HTTP, stealth browser, or anti-detect Firefox. You get clean markdown. It picks the engine.

Start scraping · $0Open the playground
POST /scrape · https://news.ycombinator.com
resolved · http · 142ms · 1 credit
HTTP
JA4 HTTP
Stealth Playwright
Camoufox
142ms
1 credit · resolved
~200ms
1 credit
~3s
3 credits
~10s
10 credits
72%of requests resolve at Tier 0
4 tierstried in order, cheapest first
1 APIone endpoint, all complexity hidden
02 · How it works

Four tiers. Cheapest first.

The router tries the fastest engine that's ever worked for a domain. If it gets blocked, it escalates. Every failure teaches it something new.

HTTP
Plain fetch · works on most static and API-backed sites
~200ms
1 credit
JA4 HTTP
curl_cffi impersonates Chrome's TLS handshake · bypasses Cloudflare without a browser
~200ms
1 credit
Stealth Playwright
Headless Chromium + fingerprint injection · for JS-heavy SPAs
~3s
3 credits
Camoufox
Patched Firefox via Python bridge · C++ level fingerprinting
~10s
10 credits
F01 · Learns
API Discovery
Browser scrapes auto-capture internal JSON endpoints and replay at 1 credit on future requests.
F02 · Race
Tier Race
Fire all engines in parallel on any URL, see which wins. Published on every intel page.
F03 · Fails loudly
Actionable errors
12 specific block reasons, each with a fix. Never get "unknown" back on a timeout.
03 · Proof

Real scrapes. Including failures.

Live pass/fail data from the production scorecard. We publish what works — and what doesn't.

17/20
Pass rate
Seed domains
142ms
Avg latency
Tier 0
$0.0001
Median cost
per scrape
72%
Tier 0 hit
No browser needed
And because it learns —
First scrape · browser
3,120ms
stealth-playwright · 3 credits
replays asdiscovered-api
Every scrape after
187ms
discovered-api · 1 credit
04 · Writing

What we've learned from 2.3M scrapes.

Production notes on anti-bot evasion, TLS fingerprinting, and engine routing. No fluff, no invented stats.

Opinion13 min read

Why 70% of Your Scraping Traffic Doesn't Need a Browser (2026)

Most scrapers default to headless Chrome for everything. We tested 60+ domains and found that roughly 70% of requests resolve at the cheapest tier — plain HTTP or JA4 fingerprinting. Here's the data, the reasoning, and what it means for how you build scraping infrastructure.

Read
Technical12 min read

Bypassing Cloudflare Turnstile in 2026: what still works

Bypass Cloudflare Turnstile in 2026: Camoufox hits 82% baseline, 54% strict. Hybrid solver patterns, named failure modes, and production limits.

Read
Technical12 min read

How to bypass DataDome in 2026 (what they're checking now)

Bypass DataDome in 2026 with Camoufox + residential proxies. ~400 behavioral signals, Poisson delays, working code, and where this stops working.

Read
All posts
04 · Pricing

You pay per credit. Not per scrape.

Tier 0 scrapes cost 1 credit. Browser scrapes cost 3. Camoufox costs 10. Plain text — no asterisks, no per-feature add-ons.

Explore

Kick the tires. No card.

$0playground
500 credits
  • 500 credits / day on the playground
  • Every tier unlocked
  • Rate limited, not feature-limited
  • No API key — browser only
Open playground
Most used

Starter

Most projects land here.

$29/ month
50,000 credits
  • 50,000 credits / month
  • At T0 that's 50k scrapes
  • API Discovery enabled
  • All 4 tiers, all engines
  • 99.5% uptime SLO
Start scraping

Scale

For production pipelines.

$149/ month
300,000 credits
  • 300,000 credits / month
  • Priority queue (p95 < 3s)
  • Dedicated router shard
  • Custom domain profiles
  • 99.9% uptime SLO
  • Email + Slack support
Talk to sales
How credits work
1 credit = 1 Tier 0 or Tier 1 scrape (plain HTTP, JA4-fingerprinted HTTP) · 3 credits = 1 Tier 2 scrape (stealth browser) · 10 credits = 1 Tier 3 scrape (Camoufox). The router always tries cheapest-first, so your real blended cost is usually close to 1.4×.