Workflow

Turn an approved script into a finished character video

Three videos went through this pipeline and were approved, including the character's debut Reel. The voice settings were the surprise - stability 0.30-0.35 plus

by Jonathan SchaferUpdated 7 Oct 2026Useful · 2

Share this workflow Post to X Post to Threads

Say this to Muse

I want you to produce a finished talking-character video from an approved script. Follow these steps in order. Ask me for anything marked [INPUT] first.
INPUTS: the approved script (final wording, don't change it); the canonical reference image for the scene (defines face, outfit, setting - never invent a different one); my ElevenLabs voice ID + API key; my HeyGen API key.
1. Validate the reference image against this checklist. Stop and report which item fails: ENTERED MY CHECKLIST HERE.
2. Generate the voiceover with ElevenLabs: MY VOICE ID, expressive inflection on closing lines.
3. Render with the HeyGen API - CONNECT YOUR HeyGEN
4. QC: face, streak, eyes...ENTER OUR LIST.  Don't pay for a re-render. If something fails unfixably, say exactly what failed and recommend: new reference image, new voiceover take, or re-render.
5. Deliver the final video plus a one-line QC summary of each check and its result.

Copy it into Muse and run it.

The result

Three videos went through this pipeline and were approved, including the character's debut Reel. The voice settings were the surprise - stability 0.30-0.35 plus explicit delivery direction turned a flat read into a great one.
What didn't work: a full-body reference image fell apart in image-to-video (face-prominent shots only now). A tailing click noise introduced in generation was fixed without a re render and by a free too muse found. Ohh, and an earlier lip-sync tool failed on consistency, which is why the checklist exists.

Questions about this workflow

What does this workflow do?

Three videos went through this pipeline and were approved, including the character's debut Reel. The voice settings were the surprise - stability 0.30-0.35 plus

What prompt does this workflow use?

I want you to produce a finished talking-character video from an approved script. Follow these steps in order. Ask me for anything marked [INPUT] first. INPUTS: the approved script (final wording, don't change it); the canonical reference image for the scene (defines face, outfit, setting - never in...

How do I use this workflow?

Copy the prompt from this page into Muse and adjust the details to your situation. The result above is what the author got. Every workflow here is free to use; you just need access to Muse, and if you do not have an invite yet you can claim a community code on the homepage.

Report this workflow

Your turn

Be one of the first to post your Muse workflow here. Add your own and become the person people follow.

Publish a workflow More workflows

Discussion

No comments yet

Be the first to share how this worked for you.