Bria FIBO 2.0: photorealistic images with reproducible JSON prompt control

Bria FIBO 2.0 turns text prompts into photorealistic images with reproducible, JSON-structured control, and returns the structured prompt and seed behind every result so you can recreate or refine it.

AMAbhilash Mirji@abhilashmirji · Oct 23, 2026 · 4 min read

Bria FIBO 2.0 is a text-to-image model from Bria that renders photorealistic images and returns the structured JSON prompt and seed behind each one. You write a short description, FIBO expands it into a detailed structured prompt, then renders from that. Because every response includes the structured prompt and the seed, you can recreate a shot exactly or change one attribute at a time. Bria trained it on licensed data.

ModelBria FIBO 2.0Generate photorealistic images with structured JSON prompt control.

What it does well

Photorealistic detail at 4MP

Switch resolution to 4MP when you need fine photoreal detail. In testing, this portrait held skin pores, deep smile lines, and individual grey beard hairs, with catchlights in the eyes and clean background bokeh. The 4MP tier adds about 30 seconds over 1MP. Use a 3:4 or 4:5 frame for head-and-shoulders shots.

Example output — Photoreal portrait at 4MPfor Bria FIBO 2.0

Reproducible, structured control

FIBO expands your prompt into a structured JSON prompt, then renders from it. Every response returns that structured prompt and the seed, so you can recreate a shot or change one attribute at a time. State object counts and positions explicitly and FIBO writes them into the structure. This prompt placed four red apples, one green pear second from the left, and a brass bell on the middle apple, in order.

Readable text in the image

FIBO can render short text inside a scene. All three menu lines here spell correctly in clean chalk, next to a small coffee-cup sketch. Put the exact words in quotation marks and keep each line short. A 3:4 or 4:5 frame gives signage room to stay legible.

Guide a look from a reference image

Pass a single reference image through image_urls and FIBO draws on its look, a mode Bria calls Inspire. Here the frosted-glass bottle and brushed-gold cap carry into a new scene — volcanic rock, drifting mist, and a rim-lit dark background — without redrawing the product. Let the reference fix the object and let the prompt describe only how the scene should change. A reference image cannot be combined with a structured prompt.

Output — Inspire mode — the reference bottle, then the new rim-lit scene FIBO rendered around it.

Input — Inspire mode — the reference bottle, then the new rim-lit scene FIBO rendered around it.

Examples to try

Each example below ran through FIBO at the settings shown on its card. Keep the returned seed and structured prompt to reproduce or refine any of them.

Name the material and the lens — "brushed titanium", "100mm macro" — and set style_id to photoreal for product-grade realism.

When hands matter, describe what they are doing, call out knuckles and fingernails, and keep the subject mid-frame so the model has room to resolve them.

Recommended settings

Setting	Value	When
resolution	4MP	Fine photoreal detail; adds about 30s. Use 1MP for fast drafts.
aspect_ratio	3:4 or 4:5	Portraits and signage. Use 16:9 for wide scenes.
style_id	photoreal	Product-grade realism; prompt-based generation only.
seed	0–2147483647	Fix it to reproduce a shot; pair it with the structured prompt.
output_type	png	Highest quality. Use jpeg for smaller files.

Good to know

A structured prompt and a reference image cannot be used together — send one or the other per request.
style_id has a single option, photoreal, and it applies to prompt-based generation only.
The 4MP tier improves fine detail but adds about 30 seconds over 1MP.
Inspire takes a single reference image (JPEG, PNG, or WEBP) through image_urls.

Use it from the API

Send a prompt and read back both the image and the structured_prompt. This is the default example's exact input.

APIGenerate from a text promptAPI reference

# pip install "segmind>=1.1.0"
# export SEGMIND_API_KEY="YOUR_API_KEY"
import segmind

result = segmind.run(
    "bria-fibo-generate",
    prompt="A white wooden windowsill in soft daylight with five red apples in a row; the second apple from the left is replaced by a single green pear; a small brass bell sits on top of the third apple from the right; sheer curtain behind, shallow depth of field",
    resolution="1MP",
    aspect_ratio="16:9",
    seed=2020,
    output_type="png",
)
print(result.get("output"))

# 1. Submit the request; the reply carries its request_id
REQUEST_ID=$(curl -s -X POST "https://api.segmind.com/v2/bria-fibo-generate" \
  -H "x-api-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "prompt": "A white wooden windowsill in soft daylight with five red apples in a row; the second apple from the left is replaced by a single green pear; a small brass bell sits on top of the third apple from the right; sheer curtain behind, shallow depth of field",
  "resolution": "1MP",
  "aspect_ratio": "16:9",
  "seed": 2020,
  "output_type": "png"
}' | jq -r .request_id)

# 2. Check its status until it is COMPLETED
curl -s "https://api.segmind.com/v2/requests/$REQUEST_ID/status" \
  -H "x-api-key: YOUR_API_KEY"

# 3. Fetch the result
curl -s "https://api.segmind.com/v2/requests/$REQUEST_ID" \
  -H "x-api-key: YOUR_API_KEY"

Resources

#text-to-image
#bria
#model-launch
#photorealistic
#structured-prompts