# Consistent Characters — workflow skill for characters

> **For AI agents:** find the right skill in one call — `curl -s "https://skills.varg.ai/api/search?q=<what the person wants>"` returns the matching use case and the best skills with install commands, as markdown (Russian works too). Full index: https://skills.varg.ai/llms.txt. Every page has a markdown twin: add `.md` to its path. Before generating anything, show the person the options on the use-case page.

Multi-scene ads where the same actors, outfits and location hold in every cut: cards, composites, Kling v3 clips.

- **Type:** [Workflow](https://skills.varg.ai/workflows.md) · [Characters](https://skills.varg.ai/browse/skills/characters.md)
- **Slug:** `varg-consistent-characters`
- **Status:** beta
- **Publisher:** varg (varg, verified)
- **License:** MIT
- **Source:** https://github.com/vargHQ/skills/tree/main/varg-consistent-characters
- **Works in:** Claude Code, Codex, Cursor, Gemini CLI, OpenClaw
- **Requires:** varg API key, varg-ai skill, ffmpeg, bun (local render)
- **Cost:** ~2,100 credits / run (estimate) — A 5-scene, 22 s ad at the per-unit prices the skill quotes (~126 credits per image, ~221 per kling_v3 clip); its older formula-care-ep2 cost table adds up to ~415. Cached re-renders cost a few credits, and an edit regenerates only the changed scene.

## What it does

Turns a scene script and real actor photos into a multi-scene vertical ad that stays consistent across cuts. It builds frozen references once (a full-body profile card per actor that fixes the outfit, and an empty location card), composes every scene from them with one shared preamble, gets the stills approved before paying for video, and animates each scene with kling_v3 native-audio dialogue in a cloud render. It also covers trimming dead air with `silencedetect`, word-level captions through a transcription pass, frame-by-frame QA, and reverse-engineering a script from a reference video. Battle-tested on a client project (formula-care-ep2).

## Use when

- A script defines actors (A1, A2…), a location (L1) and numbered scenes
- A client provides actor photos
- The same people, outfits and location must appear in every scene
- Frame-based QA of generated output
- Extracting a scene script from a reference video when none is provided

## Install

- Claude Code: `claude plugin marketplace add vargHQ/skills && claude plugin install varg@varg`
- Codex: `npx skills add vargHQ/skills --skill varg-consistent-characters -a codex -y`
- Cursor: `npx skills add vargHQ/skills --skill varg-consistent-characters -a cursor -y`
- Gemini CLI: `npx skills add vargHQ/skills --skill varg-consistent-characters -a gemini-cli -y`
- OpenClaw: `npx clawhub@latest install varg-consistent-characters`

**Models it drives:** nano_banana_pro, kling_v3, groq_whisper_large_v3_turbo

## Example output

https://skills.varg.ai/previews/varg-consistent-characters.mp4 — varg's 'Gym Ad – Two-Character UGC' template, which matches the skill's own worked example (formula-care-ep2: two actor cards, an empty gym-with-huge-windows location card, kling-v3 clips with native dialogue, 22.4 s after silence trimming). The cut (4-12 s) crosses four camera setups, and the same two actors, outfits and gym hold in every one. (credit: varg, https://varg.ai/ugc-ads)

## Used for

- [AI music video](https://skills.varg.ai/use-cases/music-video.md): A music video for your song: concept, a consistent lead character, shots cut on the beat, lip-sync on vocals.
- [AI short film or series](https://skills.varg.ai/use-cases/short-film.md): A multi-shot story with consistent characters: script, storyboard, previz, shots, edit.
- [Same character in every scene](https://skills.varg.ai/use-cases/consistent-character.md): One mascot, trainer or actor that looks identical across many images and clips.
- [Exercise demo loops](https://skills.varg.ai/use-cases/fitness-loops.md): Looping exercise demos for a fitness app: one trainer, many exercises, clean background.
- [Put your character into a reference video](https://skills.varg.ai/use-cases/character-swap.md): Your character (or person) performs the exact motion of a trending video: dance, gesture, walk, ad performance — motion transfer.

## More characters skills

- [Higgsfield Soul ID](https://skills.varg.ai/skills/higgsfield-soul-id.md): Train a reusable face model from your photos for identity-consistent images and video on Higgsfield.
- [fal Character Design](https://skills.varg.ai/skills/fal-character-design.md): Design original characters that stay consistent: reference and expression sheets, outfits, character-to-video.
- [Character Design Sheet](https://skills.varg.ai/skills/inference-sh-character-design-sheet.md): Keep a character consistent across AI images with turnarounds, expression sheets, palettes and LoRAs.