FLUX.2 JSON prompt: on Sume it is text in the prompt field
BFL says FLUX.2 accepts JSON-structured prompts. On Sume, prompt is a plain string, so JSON goes in as text; other fields are typed.

BFL says FLUX.2 supports JSON-structured prompts for precise control in production and automation. On Sume the prompt field is a plain string described as a text description of the desired image, so you can send your JSON as the string content, but Sume adds no JSON schema fields to the request. Size, count and aspect ratio stay separate typed fields.
BFL's claim is from its text-to-image page; Sume's parameters are from Image generation, read 2026-10-01.
What does BFL mean by a structured prompt?
The BFL page presents a JSON example under "Structured Prompting" and calls it ideal for production workflows and automation. This post does not restate that schema; read it on the vendor page, since BFL can change it.
How do I send one through Sume?
Serialize the object to a string and pass it as prompt with black-forest-labs/flux.2-pro. Whether the model follows your keys is a property of the model, not something Sume validates. Test with a few seeds of your own content before relying on it.
Typed request fields stay outside the string:
const structured = JSON.stringify({
scene: "ceramic mug on a wooden desk",
lighting: "soft morning window light",
});
const res = await fetch("https://api.sume.com/v1/images", {
method: "POST",
headers: {
Authorization: "Bearer " + process.env.SUME_API_KEY,
"Content-Type": "application/json",
},
body: JSON.stringify({
model: "black-forest-labs/flux.2-pro",
prompt: structured,
n: 1,
aspect_ratio: "1:1",
}),
});
console.log(res.status);Which request fields are not part of the prompt?
The parameters table lists model, prompt (string, required), n (1 to 10), resolution (512, 1K, 2K, 4K) and aspect_ratio. Anything the selected model does not list is rejected with 400 unsupported_parameter, so do not smuggle control keys outside the prompt string.
| Field | Type or range |
|---|---|
| prompt | String, required |
| n | 1 to 10 |
| resolution | 512, 1K, 2K or 4K |
| aspect_ratio | Per the model entry |
Where do I learn what works in the prompt?
Start from the FLUX.2 prompt guide, and see hex color prompts for one control that works as plain text.
Sources
Related posts
More in Developers
- FLUX.2 LoRA API: finetune_id on BFL, not on Sume
BFL serves klein LoRAs through -finetuned endpoints with finetune_id. Sume rejects provider-specific parameters, so use reference images instead.
- FLUX Request Moderated vs Content Moderated, and Sume's reason
BFL separates input moderation from output moderation and adds a Moderation Reasons array. Sume groups refusals under content_policy_rejected.
- FLUX safety_tolerance 0 to 5, and what Sume does instead
BFL's safety_tolerance runs 0 (strictest) to 5 on FLUX.2, default 2. Sume has no such dial: it returns 400 unsupported_parameter.
- Upload 30 reference images to a Sume Format run (images only)
A Format run takes up to 30 images in attachments[], type input_image, by public HTTPS URL or asset_id. Video and audio attachments are not supported today.
Written by Sume