Flux Licensing Usage Report

This is not a node you install because you want it. You install it because your Flux license asks you to - a compliance checkbox wearing a ComfyUI costume.

IG1FluxReport (you'll find it as "Flux Licensing Usage Report" under the IG1 Tools category) does exactly one job: every time the graph runs, it tells Black Forest Labs how many images you just generated with a licensed Flux model. Then it hands your image to the next node, untouched.

Who actually needs this

Let's be blunt. If you're generating with FLUX.1-dev locally under the non-commercial license - hobby use, personal art, nothing for sale - you don't need this node at all. The Dev license's whole deal is "open weights, not open source": outputs are yours to sell, hosting the model in a service is not. That bargain is the defining tension of the Flux era, and BFL polices the paid side of it with license tiers and, since 2025, increasingly tight terms.

Where this node fits is the paid side. If you hold a commercial or enterprise BFL license to run a Flux model in a product or service, usage reporting is part of that arrangement, and this node automates the part people forget. Everyone on Schnell, Klein, or an Apache-2.0 model like Qwen or Chroma: close this tab, nothing for you here.

How the reporting works

Refreshingly simple. On execute it counts the image batch - number_of_generations = len(image) in the source - and POSTs that count to https://api.bfl.ai/v1/licenses/models/{model}/usage, with your API key in an x-key header. The node's own description points to BFL's "report model usage" API reference, so this is a real, documented endpoint, not a clever hack. One catch: it's a plain HTTP call from your ComfyUI backend, so the machine running the graph needs to reach the internet.

The inputs (and the one output)

All three inputs matter, but for different reasons:

• image - your generated images. Batch size is the count. A batch of 4 counts as 4 generations, so the node is a pure passthrough that only cares how many frames flow through it.
• model - a dropdown of the licensed lineup: flux-2-dev (default), flux-1-dev, flux-1-kontext-dev, flux-tools, flux-1-krea-dev.
• api_key - the reporting key from your BFL license agreement, not your API-billing key.

The single output, IMAGE, is the untouched input. Wire it inline before a Save node and never think about it again.

Installing it

Standard dance: ComfyUI Manager → Install via Git URL → https://github.com/iguanesolutions/comfyui-flux-resolution.git, or:

cd ComfyUI/custom_nodes
git clone https://github.com/iguanesolutions/comfyui-flux-resolution

then restart ComfyUI. The pack's requirements are requests, aiohttp, torch, numpy, Pillow - ComfyUI already ships nearly all of those, and requests is the only one this node actually leans on. No model downloads.

The traps

Leave api_key empty and the node refuses to run with "The API key can not be empty" - that's the author's own validation, not a crash, but it will fail your queue. And because the report is a live network call, an offline box or a hiccup at BFL's API makes the whole run throw an exception. If you're batch-generating commercially on a machine with flaky internet, put it after the heavy work, not before it. One more: the model list is hardcoded in the pack, so brand-new licensed models only appear when the author ships an update - check the pack's repo occasionally if you're on a fresh license.