| name | sienna-nomad-anchor-payload-builder |
| description | Builds and submits ComfyUI JSON payloads for the Sienna Nomad project using pre-built Python scripts. Do NOT attempt to build payloads manually or read workflow JSON files — always use the scripts. |
Sienna Nomad — Payload Pipeline
CRITICAL RULE
Never build JSON payloads manually. Never read the workflow JSON into context.
Always use the scripts below. They handle all parsing, formatting, and submission.
Script 1 — Build all JSON payloads
When: Any time payloads need to be built or rebuilt from .md files.
python3 /Users/zgbot/Desktop/Sienna_Nomad_Project/build_payloads.py
Options:
--rebuild — overwrite all existing payloads (use after .md files change)--dry-run — show what would be built without writing anything--type anchor|photo|video — build only one type
This script scans the entire project automatically. It handles all 3 prompt formats
(standard CINEMATIC CONTEXT, markdown bullets, plain key:value). No batch files needed.
Run once, it builds everything.
Script 2 — Submit to ComfyUI and retrieve PNGs
When: ComfyUI is online and renders are needed.
Order: Always run images first, videos separately when LTX workflow is ready.
python3 /Users/zgbot/Desktop/Sienna_Nomad_Project/submit_to_comfyui.py --type image
python3 /Users/zgbot/Desktop/Sienna_Nomad_Project/submit_to_comfyui.py --type video
Options:
--resume — skip files whose PNG already exists (safe to re-run after interruption)--dry-run — list files without submitting--port 8188 — restrict to one GPU
ComfyUI endpoints
- Port 8188:
http://localhost:8188
- Port 8189:
http://localhost:8189
- Health check:
GET http://localhost:8188/api/system_stats
Project layout
- Prompts:
/Users/zgbot/Desktop/Sienna_Nomad_Project/04_Prompt_Library/
- Each episode:
EP##_Name/ with JSON_PAYLOADS/ subfolder
- Workflow template:
/Users/zgbot/workflows/hermes_z_image_turbo_api.json
- Build log:
/Users/zgbot/Desktop/Sienna_Nomad_Project/build_log.txt
- Submit log:
/Users/zgbot/Desktop/Sienna_Nomad_Project/submit_log.txt
Typical full run
python3 build_payloads.py --rebuild — build all JSONs- Start ComfyUI on Windows machine
python3 submit_to_comfyui.py --type image — render all anchors + photos- PNGs save automatically to each episode folder
