JS Rendering & Bypass

Most of the web is JavaScript. Runo handles that for you, escalating from a plain fetch to a stealth headless browser only when needed.

The Three Modes

Set via options.render_js on /extract, /batch, and /crawl:

Mode	What happens	When to use
"auto" (default)	Plain HTTP fetch first; auto-escalate to headless on signs of trouble.	The right default for almost everything.
"always"	Skip the fetch entirely; render in a stealth headless browser.	Known SPAs where the fetch will always fail. Slower.
"never"	Plain fetch only. Fail fast on JS-gated pages.	Strict cost ceiling, content known to be in the initial HTML.

The response body tells you which path served the request via render_mode:

render_mode	Meaning
"fetch"	Plain HTTP fetch was enough.
"headless"	Escalated to a stealth headless browser.
"structured"	All fields filled from page metadata.
"cache"	Result served from cache (content-type-aware TTL).

What Triggers Escalation

Under "auto", Runo escalates from fetch → headless when any of these fire:

• A known anti-bot block signature is detected.
• The body is shorter than ~500 characters.
• Visible text is sparse despite a large HTML payload (likely a JS shell).
• JS-framework markers are present.
• The HTTP response is 402, 403, 406, 429, or 503.
• Your schema asks for numbers but the page has almost no digits (weather/dashboard widgets).

Escalations are transparent. The response shape is identical, only render_mode changes.

The Bypass Ladder

Some sites push back harder than a vanilla headless browser can handle. Runo has a tiered ladder:

Tier	Mechanism	Defeats	Available
T0	Plain HTTP fetch	Static HTML	All plans
T1	TLS-impersonation client	Cloudflare passive TLS, Akamai	All plans
T2	Hardened headless browser + fingerprint bundle	CDP detection, PerimeterX, Datadome	All plans
T3	Per-host cookie persistence	Progressive-trust walls	All plans
T4	CAPTCHA solver	Interactive challenges	Pro / Scale only
T5	Residential proxy rotation	IP reputation, geofences	Pro / Scale only
T6	Archive fallback	Last resort	All plans

Per-host memory means once Runo learns a site needs T2 or T3, future calls skip the lower tiers which saves ~8 s of latency.

TIER_REQUIRED

Free and Starter plans don't include T4 (CAPTCHA solver) or T5 (residential proxies). When a host is known to require those tiers, calls on those plans return:

{
  "status": "error",
  "error": {
    "code": "TIER_REQUIRED",
    "message": "This URL requires Pro or Scale tier (CAPTCHA / residential proxy bypass).",
    "retryable": false
  }
}

HTTP 402. Upgrade or pick a different source.

Cost vs. Latency

Path	Typical Added Latency
Plain Fetch	-
Headless	~3–8 s
T4 (CAPTCHA Solver)	~5–20 s
T5 (Residential Proxy)	~1–3 s

You don't pay direct fees for T4/T5, they're included in Pro/Scale. Daily kill-switches protect against abuse spikes; if those fire, calls fall back to T6 and the response carries X-Runo-Paid-Tier-Throttled: true.