Fill in this form to immediately access the model for non commercial use

Bria AI Model weights are open source for non commercial use only, per the provided license.

FIBO-Edit 1.5 Base Hero Image

FIBO-Edit 1.5 Base extends structured image editing with multi-reference support and improved editing quality.
Built on the FIBO editing paradigm, FIBO-Edit 1.5 Base enables precise, controllable edits with support for up to four reference images for more flexible and reliable editing.

🌍 What's FIBO-Edit 1.5 Base?

Building on FIBO-Edit, FIBO-Edit 1.5 Base operates on structured JSON inputs together with one or more reference images, enabling explicit, interpretable, and repeatable editing workflows.

Version 1.5 introduces multi-reference editing with support for up to four reference images, while improving the visual quality of single-reference edits. This enables more advanced workflows such as reference-guided composition, virtual try-on, product integration, style and appearance transfer, and complex edits that require information from multiple visual inputs.

FIBO-Edit 1.5 Base is the full, non-distilled model. It is the recommended starting point for fine-tuning. The distilled, faster version is FIBO-Edit 1.5 Turbo.

Developed by Bria AI, FIBO-Edit 1.5 Base prioritizes transparency, legal safety, and granular control: ranking among the top models in open benchmarks for prompt adherence and quality.

📄 Technical report coming soon. For architecture details, see FIBO.

📐 The VGL Paradigm

FIBO-Edit 1.5 Base is natively built on Visual GenAI Language (VGL). VGL standardizes image generation by replacing vague natural language descriptions with explicit, human-machine-readable JSON. By disentangling visual elements—such as lighting, composition, style, and camera parameters—VGL transforms editing from a probabilistic guessing game into a deterministic engineering task. FIBO-Edit 1.5 Base reads these structured blueprints to perform precise updates without prompt drift, ensuring the output matches your exact specifications.

🔑 Key Features

Structured JSON Control: Move beyond "prompt drift." Define edits with explicit parameters (lighting, composition, style) using a structured JSON format for deterministic results.
Multi-Reference Editing: Supply up to four reference images in a single pass. The model draws information from every reference to compose one coherent edit, enabling reference-guided composition, virtual try-on, product integration, and style and appearance transfer.
Native Masking: Built-in support for mask-based editing allows you to target specific regions of an image with pixel-perfect precision, leaving the rest untouched.
Production-Ready Architecture: At 8B parameters, the model balances high-fidelity output with the speed and efficiency required for commercial pipelines.
Deep Customization: The lightweight architecture empowers researchers to build specialized "Edit" models for domain-specific tasks without compromising quality.
Responsible & Licensed: Trained exclusively on fully licensed data, ensuring zero copyright infringement risks for commercial users.

⚡ Quick Start

🚀 Try FIBO-Edit 1.5 Base now →

FIBO-Edit 1.5 Base is available everywhere you build, either as source-code and weights, ComfyUI nodes or API endpoints.

API Endpoint:

Source-Code & Weights

The model is open source for non-commercial use with this license
For commercial use Click here.

Installation

Install Diffusers and additional requirements:

uv pip install --torch-backend=auto git+https://github.com/huggingface/diffusers torch torchvision google-genai boltons ujson sentencepiece accelerate transformers

Usage

Write your edit instruction in plain text and convert it, together with the source image, into a structured VGL JSON prompt with briaai/FIBO-edit-gemini-prompt-to-JSON. This block calls Gemini and supports single references, multiple references, and masks. Set its API key first:

export GEMINI_API_KEY="your-api-key"

import torch
from diffusers import BriaFiboEditPipeline
from diffusers.modular_pipelines import ModularPipelineBlocks
from PIL import Image

vlm_pipe = ModularPipelineBlocks.from_pretrained(
    "briaai/FIBO-edit-gemini-prompt-to-JSON", trust_remote_code=True
)
vlm_pipe = vlm_pipe.init_pipeline()

pipe = BriaFiboEditPipeline.from_pretrained(
    "briaai/Fibo-Edit-1.5-base",
    torch_dtype=torch.bfloat16,
)
pipe.to("cuda")

# Optional: compile the repeated transformer blocks for faster inference.
# pipe.transformer.compile_repeated_blocks()

image = Image.open("source.png").convert("RGB")

# Convert the edit instruction to a structured VGL JSON prompt
json_prompt = vlm_pipe(image=image, prompt="make it look vintage").values["json_prompt"]

result = pipe(
    image=image,
    prompt=json_prompt,
    num_inference_steps=30,
    guidance_scale=5,
).images[0]
result.save("edited.png")

For mask-based editing, pass a compatible mask image to the promptify block as mask_image and to the edit pipeline as mask, along with image.

Multiple references

BriaFiboEditPipeline accepts a single source image or a list of source images. A list is treated as multiple references (not a batch): the first reference sets the output aspect ratio, and every reference is VAE-encoded at its own resolution and added as a separate conditioning stream.

references = [Image.open("ref-1.png").convert("RGB"), Image.open("ref-2.png").convert("RGB")]

json_prompt = vlm_pipe(
    image=references, prompt="place the subject from both references on a beach"
).values["json_prompt"]

result = pipe(
    image=references,
    prompt=json_prompt,
    num_inference_steps=30,
    guidance_scale=5.0,
).images[0]
result.save("multi_ref.png")

More Examples

Input	Output

If you have questions about this repository, feedback to share, or want to contribute directly, we welcome your issues and pull requests on GitHub. Your contributions help make FIBO better for everyone.

If you're passionate about fundamental research, we're hiring full-time employees (FTEs) and research interns. Don't wait - reach out to us at hr@bria.ai

Citation

@article{gutflaish2025generating,
  title={Generating an Image From 1,000 Words: Enhancing Text-to-Image With Structured Captions},
  author={Gutflaish, Eyal and Kachlon, Eliran and Zisman, Hezi and Hacham, Tal and Sarid, Nimrod and Visheratin, Alexander and Huberman, Saar and Davidi, Gal and Bukchin, Guy and Goldberg, Kfir and others},
  journal={arXiv preprint arXiv:2511.06876},
  year={2025}
}