voiceover — Claude Code command: what it does, how to install
Runs demo-driven feature workflows by agreeing on a voiceover script before opening a git worktree, building code, and filing a PR.
LicenseRef-Proprietary ★ 7,523 updated 2026-10-29
.claude/commands/voiceover.mdWhat voiceover does
This command swaps out standard PRDs for a narrator script describing an end-user walkthrough. The agent drafts spoken narration frame by frame before touching any code. If an action sounds awkward when spoken aloud, the design gets revised immediately. Coding stays blocked until the draft receives explicit approval.
Once confirmed, the process branches off into a separate directory via git worktree based on dev. It stores the narration markdown in evals/voiceovers/<flow-id>.md and initializes files using pnpm fraimz scaffold. An executor subagent handles implementation while the fraimz skill runs validation checks on each frame. The flow finishes with gh pr create and proof reporting. The workflow fails immediately if fraimz or gh are absent from the shell environment.
What it helps with
- draft demo narration scripts to define requirements before coding
- spin up isolated branches using git worktree
- scaffold testable feature flows via pnpm fraimz
- open pull requests with verification proofs attached
Example prompts for the agent
- /voiceover Add CSV export for monthly billing reports
- /voiceover Implement one-time email passcode login
How to install voiceover
Manually: put the files in place
.claude/commands/voiceover.mdCall it with /voiceover. For all projects use ~/.claude/commands/
Manually: put the files in place
~/.codex/prompts/voiceover.mdCall it with /prompts:voiceover. Codex marks this format deprecated in favour of skills
Manually: put the files in place
.github/prompts/voiceover.prompt.mdManually: put the files in place
.gemini/commands/voiceover.tomlGemini stores commands as TOML: the prompt text goes into the prompt field
Questions about voiceover
How to install voiceover in Claude Code?
Save the command instructions into .claude/commands/voiceover.md inside your project directory.
What does the voiceover command do?
It guides feature work through demo-first specs. You review the user narration, scaffold the files in an isolated worktree, and test every frame before pushing a pull request.
What dependencies does voiceover expect?
The workflow relies on Git worktrees, the gh command-line tool, and pnpm fraimz for scaffolding and frame verification.