How do I get the best identity match from the headshot?

Use a high-resolution, front-facing headshot with neutral lighting. Load the headshot first as the primary reference. Keep prompts consistent between the still and video steps, and avoid conflicting descriptors (e.g., changing hair color) unless intentional.

Can I change the wardrobe, product, or environment?

Yes. Update the INPUT SCENE DESCRIPTION text and, if available, provide product/environment images via LoadImage. The StringConcatenate node merges these details into the final prompt used by GeminiImage2Node, which will guide the look in the still and carry through to the video.

What controls the camera movement and action in the video?

Your motion cues in the INPUT SCENE DESCRIPTION for STEP 2. Describe actions concisely (e.g., 'talent takes a sip, smiles, then glances at the can; slow 2–3s dolly-in'). KlingOmniProImageToVideoNode uses the approved still plus these cues to generate a short motion pass.

Where do I find the outputs, and what formats are saved?

SaveImage writes the stills to the ComfyUI/output images folder, and SaveVideo writes the motion previews to the ComfyUI/output videos folder. Filenames and formats follow your node settings; check the SaveImage/SaveVideo nodes to adjust naming and codec.

Comfy for Advertising - Talent Casting

This workflow builds fast, client-ready casting previs for advertising in two passes: a photo-real still and a short motion test. In STEP 1 – PREVIS STILL FRAME, your talent headshot (LoadImage) is combined with a scene description to synthesize a single frame of the talent in context (e.g., tennis court, soda product). Prompt text from MarkdownNote and PrimitiveNode is merged with shot directions using StringConcatenate, and GeminiImage2Node performs image-to-image guided synthesis using the headshot and any optional product/environment references. The resulting frame is saved with SaveImage for quick side-by-side casting reviews.

STEP 2 – PREVIS VIDEO upgrades that still into a short motion clip. ImageBatchMulti assembles one or more selected stills into a batch, and KlingOmniProImageToVideoNode generates a quick image-to-video preview—ideal for validating basic motion like a gentle dolly-in, a sip, and a look-to-can. SaveVideo writes the final clip to disk. Because the video is derived from your approved still, character identity and wardrobe cues tend to remain consistent across frames, making it well-suited for talent casting, pitch decks, and early client sign-off.

API

Use this workflow from code

Every Comfy workflow is a JSON graph. The payload below is this workflow, exactly as ComfyUI runs it — fetch it from the URL, keep it in version control, or load it in ComfyUI and run it node by node.

ff88e6723334.json

Fetching workflow JSON…

Run it from TypeScript or Python with the Comfy SDK. The same code targets Comfy Cloud or a ComfyUI you host yourself — only the base URL changes.

// Install (beta)
npm i @comfyorg/sdk

// Run this workflow (TypeScript)
import { Comfy } from "@comfyorg/sdk";

const client = new Comfy({ apiKey: "comfyui-..." });
const wf = await client.workflows.fromFile("workflow_api.json");
const job = await client.run(wf);
await job.getOutputs("<output-node-id>")[0].toFile("output.png");

The SDK takes a workflow in API format: open this workflow in ComfyUI and use File → Export Workflow (API).

Comfy Cloud API access requires a plan with an API key.

SDK docs Get an API key

FAQ

Frequently Asked Questions