Limits & Timeouts¶

Every HTTP request operates inside guardrails — timeouts, response sizes, redirect caps, stream lifetimes. Values outside allowed ranges are clamped automatically. You won't get a mysterious crash; you'll get a polite res.ok === false.

This page is the cheat sheet. Bookmark it for when something times out and you're not sure why.

Request timeout¶

// Uses plan default
await HTTP.get("https://api.example.com/data")

// Custom (clamped to plan max)
await HTTP.get("https://api.example.com/data", { timeout: 5000 })

On timeout: res.ok === false, res.status === 408, res.error.code === "TIMEOUT".

Response size¶

Responses larger than the limit return res.ok === false with error.code of "RESPONSE_SIZE_EXCEEDED" or "RESPONSE_TOO_LARGE".

For very large payloads, use responseType: "stream" and read incrementally — see Streaming. Don't try to swallow a whale in one bite.

Redirects¶

Followed status codes: 301, 302, 303, 307, 308.

When the redirect limit is exceeded: res.error.code === "REDIRECT_LIMIT_EXCEEDED", res.status === 310.

await HTTP.get(url, { maxRedirect: 5 })   // up to 5
await HTTP.get(url, { maxRedirect: 0 })   // no redirects
await HTTP.get(url, { redirect: false })  // disable following

Streaming limits¶

Full guide: Streaming. Summary from the implementation:

Stream lifetime by plan¶

Streams are cancelled when lifetime or idle timeout is exceeded.

Command chain limit¶

Each success or error HTTP callback counts toward the 6 nested command limit per execution (shared with Bot.run).

Proxy¶

Cloudflare Worker proxy¶

Validation limits¶

Invalid options throw before the request is sent. Catches typos early.

Quick reference¶

See your full plan details on the plan global.