Errors.

{
  "type": "https://snakereach.com/errors/authentication-required",
  "title": "Authentication required",
  "status": 401,
  "detail": "Provide your key in the X-API-Key header, or as Authorization: Bearer <key>."
}

title is the machine-stable summary and detail is the human one. Read both. Where a status has several causes, title is what distinguishes them.

Every status

• 400 · Malformed requestThe request body was not valid JSON.
• 401 · Authentication requiredNo API key was sent.
• 403 · ForbiddenThe key was rejected, robots.txt disallowed the URL, or it resolves to a non-public address.
• 404 · Job not foundNo crawl job with that id.
• 413 · Response too largeThe upstream page exceeded the response size cap.
• 415 · Unsupported content typeNo parser is installed for that content type.
• 422 · Invalid fieldThe body was valid JSON but a field was wrong.
• 429 · Rate limit exceededToo many requests, too many concurrent crawls, or a budget was exhausted.
• 502 · Upstream fetch failedThe page could not be fetched at all.
• 504 · Upstream timeoutThe target site did not respond in time.

What is not an error

A target site returning 404, 403 or 500 is not an API error. You get a normal Document with that status and whatever content the page had. The API only errors when it could not obtain a response at all, or when policy refused to try.