# lllyasviel/Omost

Your image is almost there!

Repository: https://github.com/lllyasviel/Omost
Canonical: https://ross.abutalabs.com/products/omost
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2024-07-26T21:10:06+00:00

## Health v2 (maintenance only)
Score: 24/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 59
- inputs: {"age_days": 826, "days_push": 768, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 7610, forks 434 (observed 2026-08-28T04:10:02.010775+00:00)

## What it is
Omost is a Python application that converts LLM coding capability into image composition by having pretrained LLMs (based on Llama3 and Phi3) write code against a virtual Canvas agent. The Canvas is rendered by image generators to produce final images, runnable locally with 8GB Nvidia VRAM or via a hosted HuggingFace space.

## Use cases
- generate images from text prompts with better layout control
- compose complex multi-subject scenes for image generation
- use an LLM to plan image content before diffusion rendering
- run a local text-to-image assistant on a consumer GPU
- improve prompt-to-image composition quality

## When to choose
- you want more controllable, composed image generation than plain prompt-to-image tools
- you have an Nvidia GPU with at least 8GB VRAM and want a local Gradio app
- you want to experiment with LLM-driven image layout planning

## When to avoid
- you need a production-grade or API-first image generation service
- you have no GPU and cannot rely on the hosted HuggingFace space
- you need fine-tuning or training of the underlying models

## Facets
- artifact type: application
- maturity: active
- function: llm-inference, image-processing, machine-learning, agent-framework
- domain: artificial-intelligence, large-language-models, image-processing, graphics
- platform: python, self-hosted
- tags: image-generation, text-to-image, gradio, diffusion, canvas-composition, llm, gpu, web-server

## Member repositories
- lllyasviel/Omost (main) score 24

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:10:02.010775+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:36:50.431552+00:00, confidence not recorded.
  - readme: https://github.com/lllyasviel/Omost (fetched 2026-08-28T04:10:02.010775+00:00, sha 6e6c2951ada5)
- Data as of 2026-08-30T08:39:29.467469+00:00.