📷 Passport Prompt

Here's a sentence you don't see every day on comfy.icu: this node exists to make your face acceptable to the United States government. Passport Prompt outputs a ready-made instruction for an image-editing model - "Make a professional USA passport photo: pure white background, center the face and shoulders perfectly, neutral expression with both eyes open and mouth closed, even studio lighting with no shadows, head occupies 50-69% of image height" - so you can turn a casual selfie into something a passport office won't bounce.

It's the odd node out in the comfyui-auto-anime-prompts pack, which is otherwise all anime tag generation. The author added it alongside a Qwen-based passport workflow, and it's clearly aimed at Qwen-Image-Edit-2511 - the open Apache-2.0 editor model where you hand it a picture and a sentence and it just does the edit, no mask, no ControlNet.

How it works

Boring in the best way: it returns a hardcoded prompt string (and an empty negative, because Qwen-class editors work best with the negative box empty - garbage in the negative is how you get weird artifacts). That's the entire mechanism, and it's why the node is worth keeping around: the passport specs are exact (head 50–69% of image height, neutral expression, eyes open, mouth closed), and the author already did the research on the compliance wording so you don't have to.

Three inputs, all simple:

• use_default_prompt (on) - use that compliance prompt above.
• custom_prompt - replace it entirely with your own wording.
• append_to_default - bolt extra instructions onto the end ("keep my glasses", "no headscarf") while keeping the compliance core. This is the one you'll actually touch.

Outputs: prompt and negative. Wire prompt into your CLIP Text Encode (Positive Prompt) node. In the author's own my-qwen2511-edit-image-v1.json workflow that's a Qwen-Image-Edit 2511 setup: 8 steps, CFG 1, euler/simple, a Lightning LoRA, and a Qwen2.5-VL CLIP loaded as GGUF - the standard low-VRAM path for this model.

Install

ComfyUI Manager: search "Auto Anime Prompt", or:

cd ComfyUI/custom_nodes
git clone https://github.com/jluo-github/comfyui-auto-anime-prompts

No extra dependencies, no model downloads from this pack itself - the model (Qwen-Image-Edit + its CLIP GGUF) is on you, and it's the actual heavy lift. Restart ComfyUI after cloning.

One honest caveat

The node gives you a prompt, not a compliant photo. Qwen-Image-Edit re-emits the whole frame, and its known limitations - geometric offset, face drift - apply exactly where passport rules are strictest. The output may still need a resize and a crop to meet the 2x2 spec, which is why the pack pairs this node with its two siblings, Passport Resize and Passport Tile. Use all three together and the pipeline is genuinely coherent. And, you know: check with the passport office before you rely on an AI edit for a legal document.