兔子API 兔子API
中

provider

Black Forest Labs

Direct answerFocused on high-quality image generation and editing across realtime previews, production workflows, and fine-grained control.

Updated · Reviewed

Separate BFL-native jobs from this site's image API

Black Forest Labs is the originator of FLUX image models. Direct BFL access is an asynchronous job protocol authenticated with x-key. This site's /v1/images/generations is an adapted gateway route using this site's Bearer token; BFL /v1/flux-* and get_result paths must not be presented as routes here. Confirm the exact model, editing mode, dimensions, and reference count per model page.

Persist the returned polling_url

A successful submission returns a job id and polling_url. With global api.bfl.ai or multi-cluster regional endpoints, poll the exact returned URL instead of constructing a hostname from an older region. Pending means wait, and Ready exposes result.sample. Handle Request Moderated, Content Moderated, Task not found, and Error separately. Use backoff, a total deadline, and cancellation.

Result URLs last 10 minutes and have no CORS

Signed delivery.*.bfl.ai URLs expire after about 10 minutes and do not permit browser cross-origin access. After Ready, stream-download the asset server-side into controlled object storage or a CDN and return your own durable URL. Do not persist the delivery URL as a permanent asset or make browser rendering depend on it.

402 and 429 require different actions

402 Payment Required indicates insufficient credits or a payment-method issue and will not recover through retry. 429 Too Many Requests indicates a rate or active-job limit; wait for work to finish and lower concurrency. Fix 400/422 inputs, check key permission for 403, and retry only transient 500/503 failures with bounded jittered backoff. Keep a business idempotency record because duplicate submissions can be billed twice.

Licensing follows the exact product

BFL's hosted API is governed by API service terms and usage policy, while downloadable weights carry their own licenses. Some open weights use Apache 2.0; others are non-commercial or require separate commercial rights—for example, commercial use of Kontext [dev] requires licensing. Record the exact model, weight source, license version, input rights, and moderation requirements before launch.

Reviewed provider catalog

Latest models by capability

These are provider-catalog models, not a promise of availability on this site. Confirm callable IDs, endpoints, and pricing in the model marketplace.

Image generation and editing

Check live availability in the model marketplace

Use cases

  • Image generation and editing
  • High-quality production output
  • Typography and detail control

API protocols

  • /v1/images/generations

FAQ

Should a BFL API key be sent as Bearer?

No. Direct FLUX API calls use x-key: <BFL_API_KEY>. This site's /v1/images/generations uses this site's Bearer token; the credentials are not interchangeable.

Which result URL should I construct after creating a FLUX job?

Do not construct one. Persist both id and the returned polling_url, then request that exact URL. Global and multi-cluster regional endpoints explicitly require it.

Why can a browser not keep displaying a BFL result?

result.sample is a signed URL valid for about 10 minutes, and delivery URLs do not enable CORS. Download it server-side into your object storage or CDN immediately.

What is the difference between BFL errors 402 and 429?

402 means credits or payment need attention; add funds before submitting again. 429 means an active-job or rate limit, so reduce concurrency and wait for existing work.

Can every open FLUX weight be used commercially?

Not automatically. Hosted API terms and weight licenses are separate. Exact FLUX releases can use Apache 2.0, a non-commercial license, or require separate commercial licensing.

  • FLUX image modelsFLUX.2 officially separates klein, max, pro, and flex for fast local inference, maximum consistency, production quality, and typography, with different APIs and licenses.
  • Image generation and vision model guideVision understanding and image generation use different interfaces; generation projects must validate prompt adherence, editing consistency, text, likeness rights, moderation, and output terms.
  • AI API quickstart: from API key to first requestAn AI API call needs a matching base URL, API key, model ID, and endpoint; start with minimal curl, then add error handling, cost controls, and key security.
  • Common AI API error codes and troubleshootingHandle AI API failures by parsing the HTTP status, provider code, error type, message, retry hints, and request ID together. Fix request, authentication, billing, and hard-quota causes first; retry only transient throttling, network failures, and selected 5xx errors with strict bounds.

Official sources

  1. BFL Image Generation Quickstart Official
  2. FLUX API Integration Guide Official
  3. FLUX API Errors Official
  4. BFL API Pricing Official
  5. FLUX Kontext Licensing Overview Official