Automated long-form AI visual generation pipeline for music, DJ sets, and experimental audiovisual projects.
Runs a sliding-window paragraph story through a configurable backend chain — generating images, videos, and morphs per cycle — and assembles them into a final output video.
Backends: ComfyUI · Venice · Veo
Features: Reinject / img2vid chaining · Crowd Control QR overlay · Live OBS integration · ETA spinner · Per-op timing
# 1. Clone & install
git clone https://your-repo/visual-faqtory.git
cd visual-faqtory
pip install -r requirements.txt
# 2. Set API key (Venice example)
export VENICE_API_KEY=your_key_here
# 3. Drop your story into worqspace/
cp my_story.txt worqspace/story.txt
# 4. Optionally add a base image
cp my_base.jpg worqspace/base_images/
# 5. Run
python vfaq_cli.py run -n my-run
# 6. Resume from checkpoint
python vfaq_cli.py run -n my-run --resumeConfig is in worqspace/config.yaml. The default is Venice all-backends.
See doc/DOCUMENTATION.md for full config reference.
worqspace/ User workspace — story, config, base images, prompts
vfaq/ Core Python package
venice_backend.py Venice image + video backend
veo_backend.py Google Veo backend
backends.py Backend interface + ComfyUI backend
sliding_story_engine.py Main cycle engine
visual_faqtory.py Run orchestrator
crowd_control/ Live QR crowd-control server
vfaq_cli.py CLI entry point
doc/ Documentation
| Doc | Description |
|---|---|
doc/DOCUMENTATION.md |
Full config reference, pipeline architecture, backend guide |
doc/LIVE-INTEGRATION-GUIDE.md |
OBS + SRT live streaming integration |
doc/EXTERNAL-LIVE-VISUALS-SETUP.md |
External live visuals setup |
doc/RELEASE-NOTES.md |
Full changelog |
doc/BACKEND-VALIDATION-REPORT.md |
Backend validation results |
doc/VALIDATION-REPORT.md |
Pipeline validation report |
doc/SRT-LIVE-OPS-REPORT.md |
SRT live ops report |
Visual FaQtory is licensed under the GNU Affero General Public License v3.0 (AGPL-3.0). See the LICENSE file for full text.
