Skip to content

README.md

Latest commit

 

History

History
104 lines (78 loc) · 3.6 KB

README.md

File metadata and controls

104 lines (78 loc) · 3.6 KB

FrameFamily

A modular AI ecosystem focused on frame-based image generation, training, and visualization.

Components

  • Training
    FrameForge
    AI training, dataset preparation, and orchestration within the Frame ecosystem.

  • Viewing
    FrameView
    Visualization, inspection, and analysis of generated frames and training results.

  • Generating
    FrameCreate
    Generative image AI of the Frame ecosystem.
    Work in Progress

FrameCreate

License Node Postgres

FrameCreate is the generative core of the FrameFamily. It gives you a clean, fast image generator with a calm UI, model control, and a clear history of every output.

Work in Progress

Notice: Right now only SDXL-based models are supported. Embeddings are not wired yet.

FrameCreate is built for creative, synthetic, and stylized content. Use on real individuals without consent is explicitly discouraged.

Support and Questions -> Discord
https://discord.gg/TB5DHMNa5J

Why you might like it

  • One place to generate, manage models, and review results.
  • A clear, uncluttered workflow that stays consistent with FrameFamily.
  • Fast queue handling so the machine stays focused on generation.
  • Built to stay fully open and self-hosted.

What you can do

  • Generate images with live preview and stop running jobs when needed.
  • Manage base models, LoRAs, and VAEs in one place.
  • Stack up to three LoRAs and control each strength.
  • Browse history with metadata, reuse prompts, and delete what you do not need.
  • Use preset styles and wildcard prompts to speed up prompting.
  • Set default sampling and live preview settings in System.

Quick Start

./scripts/setup.sh

Open the Web UI at http://localhost:5174. The setup script installs dependencies, prepares the database, runs migrations, and enables systemd services.

What you need: Node.js + npm, Python 3, and Postgres. A GPU is recommended for generation.

First steps

  1. Run the setup command above.
  2. Open the web UI.
  3. Drop your models into the storage/ folders (see below).
  4. Use the Model Manager to rescan.
  5. Generate your first image.

Storage Layout

FrameCreate stores everything it needs under the storage/ folder. You can drop your models there and FrameCreate will find them.

storage/
  models/       # base checkpoints (.safetensors)
  loras/        # LoRA adapters (.safetensors)
  vaes/         # VAE weights
  embeddings/   # text embeddings
  outputs/      # generated images
  thumbnails/   # UI thumbnails
  wildcards/    # prompt wildcard lists (.txt)

Tip: After adding models, open the Model Manager and click Rescan.

Wildcard prompts

Drop a text file into storage/wildcards/. Each line is one option. Use it in your prompt like __colors__.

Example:

  • storage/wildcards/colors.txt 
    red
blue
green
  • Prompt: a __colors__ car
    Each image in a series uses the next line from the file; when the batch exceeds the list, values cycle from the top. Lines without letters are ignored.

Optional: send wildcard_strategy in the job request (sequential, cycle, random) to control selection.

Optional: advanced setup

If you want to change ports, database settings, or runtime options, edit .env. You can start from .env.example.

License

MIT