Flux Finetune Details (BFL)

Flux Finetune Details (BFL) is the drill-down counterpart to Flux My Finetunes in the ComfyUI-FLUX-BFL-API pack. Where "My Finetunes" lists everything on your account, this node takes a single finetune_id and returns that one model's full details as JSON - the metadata BFL keeps about the finetune: its ID, status, configuration, and whatever else the finetune_details endpoint returns for a given custom model.

So when would you actually reach for it? Mostly diagnosis and bookkeeping. You've trained a finetune, it's behaving oddly, and you want to see its exact stored configuration. Or you grabbed an ID from Flux My Finetunes and want to double-check which model that ID actually refers to before you burn a generation on it. It's the "inspect before you use" step in the pack's finetune workflow.

How it works

It's a GET to the finetune_details endpoint with finetune_id as the query parameter. The response JSON comes out the STRING output, pretty-printed. No images, no generation - pure account lookup. Same shape as the other finetune-management nodes: one ID in, JSON out.

The inputs and output

• finetune_id - the ID of the finetune to inspect. Grab it from Flux My Finetunes or your account dashboard.
• config (optional) - a Flux Config (BFL) node. And yes, the region rule applies here too: the default api.bfl.ai/v1/ doesn't serve finetune routes, so region must be set to us or eu for this call to work.
• Output: STRING - the JSON details blob.

Installing it

Same pack, same one-time install:

cd ComfyUI/custom_nodes
git clone https://github.com/gelasdev/ComfyUI-FLUX-BFL-API.git

Restart, key in config.ini or a Flux Config (BFL) node, region set. torch is the only declared dependency.

Where people get burned

• The regional endpoint, for the third time. It's worth saying plainly because it's the single most common finetune failure in this pack: without region set on the config node, every finetune call - details included - errors out. Get this right once and the whole finetune section works.
• Expecting an image. This is a metadata node. The "result" you might want lives in the status/output of the actual finetuned generation, not here.
• Blanking on which ID to use. If you're inspecting a finetune you trained, use its training ID. If you're inspecting a generation run that used a finetune, that's a task ID - different endpoint, different meaning.

It's a quiet node, but it rounds out a real workflow: My Finetunes to find IDs, Details to confirm which model an ID is, Status to track the run, Delete to clean up. For anyone managing more than a couple of custom models on BFL, that loop is the pack's underrated value.