Extensions/ꘇ Genereti
ComfyUI Extension

ꘇ Genereti

EARLY ALPHA - experimental classroom release. Experimental realtime live creative tools for ComfyUI: code (p5js,glsl), drawing, shaders, 2d operators, web audio, MIDI, lessons and performance timelines. Expect bugs and changing interfaces; back up workflows.

By languel·Created 8 days ago·Updated about 17 hours ago· 1
languel/genereti
Nodes—
On cloudLocal install
Stars1
Updatedabout 17 hours ago
Readme

Genereti

Realtime creative coding and performance tools for ComfyUI. Build live visuals and sound with code, drawing, shaders, texture operators, signals, interactive lessons and timeline automation.

Super alpha — early classroom testing. Back up your workflows before updating. Expect bugs and changing interfaces.

Student getting started

The bundled demos need no model downloads, separate model server, or extra node packs. Use a recent ComfyUI installation and a modern browser; GPU texture demos require WebGPU support.

  1. Install ꘇ Genereti through Comfy's Extensions / Manager, then restart and refresh the browser. For a GitHub installation, stop Comfy and run this from your ComfyUI directory:

    git clone https://github.com/languel/genereti.git custom_nodes/genereti

  2. For GitHub installs, install the requirements using ComfyUI's Python. On a checkout with a .venv:

    ./.venv/bin/python -m pip install -r custom_nodes/genereti/requirements.txt

    See the installation guide for Windows, Portable, Desktop and existing installations.

  3. Restart ComfyUI and refresh its browser page.

  4. Open Workflows → Genereti or Templates → genereti. Start with ꘇ-Performance-Timeline; Alt+Shift+T opens the timeline. Audio starts explicitly on mod.output.

Continue with the student quickstart, shader, p5 and audiovisual tutorials, or demo list.

What is included

  • Livecode: p5, GLSL/multibuffer shaders, Three.js, HTML and document rendering, with editable parameters.
  • Drawing and capture: interactive Excalidraw, explicitly started camera/screen sources, and image output views.
  • OpenTouch: top.* textures, chop.* signals and dat.* data/lessons.
  • Sound: mod.* browser instruments, effects, analysis and visualizers.
  • Performance: shared time, musical timing, scale, automation and a docked timeline.
  • Assistant: optional local/hosted provider connections; no assistant service is needed for the bundled demos.

See the node catalog, audio guide, performance guide and keyboard shortcuts.

AI inference is moving to GeneretiCore

Model generation, conversion and downloads belong in the separate experimental GeneretiCore repository. It is not yet an installable replacement. Genereti's current classroom demos use the lightweight performance framework only.

Old standalone apps, inference packs, conversion scripts and model-dependent demos are preserved in the migration archive. They are unsupported here and excluded from the distributed package.

For contributors

  • integrations/: the supported Comfy node packs; the legacy archive is clearly separated.
  • example_workflows/: bundled templates shown by Comfy.
  • docs/: guides and design notes.
  • editor/ and web/vendor/: drawing editor source and its shared browser bundle.
  • scripts/, tests/ and distribution/: developer build, validation and packaging tools.

Canonical workflow sources remain in integrations/comfyui_genereti/workflows/; run python scripts/sync_comfy_templates.py after editing them. Students do not need npm or these build scripts.

Original Genereti code is MIT licensed. Dependencies retain their own licenses and notices; see licensing. Model weights, secrets and generated captures are not stored in this repository.