Stable Yogi
d5ddab3a5f
Same as the Forge extension: the node already fetches one prompt per queue (1:1 with allowance), but the README/tooltip still described the old local-cache behavior. All user-facing text now says it plainly: 1 queue = 1 call = 1 prompt, nothing stored.
StableYogi Prompt — ComfyUI node
Adds a StableYogi Prompt node that outputs a ready-made prompt string. Wire it into CLIP Text Encode and every queue pulls a fresh prompt.
Built by stableyogi.com. Requires an account key (Settings → Prompt Engine API).
Install
- Copy this folder into
ComfyUI/custom_nodes/. - Restart ComfyUI.
- Add the StableYogi Prompt node (category StableYogi), paste your key into api_key once (it's saved locally, then you can clear it), pick a subject, and connect prompt → CLIP Text Encode.
Inputs
- subject — who the prompt is about.
- trigger — LoRA / character trigger word, prepended.
- quality — quality suffix.
- style —
nl(sentence) ortags. - api_key (optional) — paste once; kept in a local file, not in your saved workflow.
- refresh (optional) — resets the node's status; every queue already fetches fresh (nothing is stored).
How your allowance is spent (simple + fair)
Every queued run makes exactly one API call for exactly one prompt: 1 queue = 1 prompt = 1 from your allowance. Nothing is pre-fetched, batched, or stored on your machine — you only ever spend what you actually use.
Notes
- Each queue outputs a fresh, varied prompt — fetched on demand at queue time.
- Your key is stored locally and is not written into the workflow when the field is left empty.
- No extra packages required.
More free tools by Stable Yogi
Small, free, open tools for local AI art — Forge / Forge Neo, AUTOMATIC1111, and ComfyUI. Browse them all at github.com/Stable-yogi · more at stableyogi.com.