narrator — Codex skill: narration and installation
narrator creates voice audio for timed video segments or a continuous long-form story. Use it when consistent voice and checked pauses matter.
npx skills add openai/plugins --skill narrator -a codexWhat narrator does
For video segments, narrator generates a separate audio take for each line and checks spoken duration and internal pauses. When a take misses its window, the skill revises the wording and generates it again; takes that pass are left untouched. It does not stretch or speed up audio.
For a continuous story, the text can be split into large chunks and joined in reading order. The workflow needs the lines to speak and a voice_id plus voice_type; in Codex it uses Higgsfield voice tools. If the text or voice pair is missing, ask for it. Ordinary read-aloud requests, voice cloning, and dubbing existing speech are outside its scope.
What it helps with
- Fit each narration line to a video segment window
- Regenerate takes that miss the spoken-duration check
- Join chunks into one long-form narration file
- Keep the same voice across audio segments
Example prompts for the agent
- Use the narrator skill to turn these numbered lines into takes that fit their video windows.
- Run narrator for one continuous story with
voice_idandvoice_type. - Check spoken duration and pauses for each line, then regenerate only the takes that fail.
How to install narrator
1Skills CLI (Vercel)
npx skills add openai/plugins --skill narrator -a codex2Manually: put the files in place
.agents/skills/narrator/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add openai/plugins --skill narrator -a claude-code2Manually: put the files in place
~/.claude/skills/narrator/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/narrator/
Manually: put the files in place
.cursor/skills/narrator/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install openai/plugins narrator2Manually: put the files in place
.github/skills/narrator/SKILL.mdPlatform CLI
gemini skills install https://github.com/openai/plugins --path plugins/higgsfield/skills/narratorManually: put the files in place
.devin/skills/narrator/SKILL.mdThe legacy .windsurf/skills/ path still works
Claude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: Codex skills: how to install, where they live, how to call them
Questions about narrator
How do I install narrator in Codex?
Run npx skills add openai/plugins --skill narrator. To specify Codex explicitly, use npx skills add openai/plugins --skill narrator -a codex.
What does narrator do?
It generates audio either as timed takes for separate lines or as a continuous reading of a longer script. In line mode, it checks spoken duration and internal pauses.
What does narrator need to run?
It needs the narration text and a chosen voice_id and voice_type pair. Checking downloaded takes against video windows requires audio access in sandbox_exec.
