agentleFS
Sign inStart free
SentryKnown publishergetsentry/junior/packages/junior-tts/skills/tts/SKILL.md

tts

Skill394 starsChanged 7 months agoApache-2.0View on GitHub

Generates spoken narration with Gemini TTS through Vercel AI Gateway. Use for text-to-speech, audio samples, voiceovers, or speaking as Junior with the reusable Junior voice. Do not use for transcription, voice cloning, music generation, or rendering a video without a narration request.

  • Reads credentials

We split this skill into the rules it gives an agent

5 rules. 0 are common, 1 is distinctive, and 4 are very specific to this project.

  • 0

    Common

    Most of the 17 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Use the bundled scripts/speech.mjs from this directory.

    line 8
  2. Very specific:

    Use the default junior-v1 preset: a bright, warm, mischievous adult helper with dry humor.

    line 13
  3. Distinctive:

    Write narration to a UTF-8 file.

    line 14
  4. Very specific:

    Use sendFiles to deliver the returned path as audio/wav.

    line 16
  5. Very specific:

    Read assets/junior-v1.json when the user asks how the voice is defined.

    line 34

“Common” is measured: kept by at least 6 in 10 of the 17 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Text to Speech

Use the bundled scripts/speech.mjs from this directory. Node.js and the installed plugin's host-managed Gateway key are required. Do not install an SDK or ask for secrets in chat.

Workflow

  1. Resolve the requested spoken text. For a sample of Junior's voice, use the bundled sample. Do not send unrelated private context or the full persona to the provider.
  2. Use the default junior-v1 preset: a bright, warm, mischievous adult helper with dry humor. Use --delivery plain for serious or sensitive text. Do not present it as an imitation or recording of a real performer.
  3. Write narration to a UTF-8 file. Keep each clip at 1–5000 characters. Generation spends Gateway credits. Do not generate extra variants without a request.
  4. Generate to a fresh .wav path. Inspect an existing file before trying again. Do not automatically retry timeouts or failures: generation may already have spent credits.
  5. Use sendFiles to deliver the returned path as audio/wav. Say it is synthetic narration. For a video, measure the new audio and pass it to the rendering workflow; speech generation alone does not update the video.

Commands and Assets

# Local, no API call: show the reusable voice preset.
node scripts/speech.mjs --profile

# Paid: generate the bundled Junior sample.
node scripts/speech.mjs --sample --output /tmp/junior-sample.wav

# Paid: generate requested narration.
node scripts/speech.mjs --text-file /tmp/narration.txt --output /tmp/narration.wav

# Paid: use the same voice without snark.
node scripts/speech.mjs --text-file /tmp/incident.txt --output /tmp/incident.wav --delivery plain

Read assets/junior-v1.json when the user asks how the voice is defined. Read assets/sample.txt when previewing the sample script. The script sends delivery instructions separately from text. Do not prepend those directions to narration.

Results are JSON with path, bytes, MIME type, model, voice, profile, delivery, and synthetic status. Use --help for the command surface. No key argument is accepted.

Failures

  • Missing host key or HTTP 401/403: ask the operator to check AI_GATEWAY_API_KEY, plugin registration, and team speech access, then redeploy. A missing key in the Sandbox is expected. Core's OIDC authentication is not reused by this manifest-only plugin.
  • HTTP 402: ask the operator to check credits and spending limits.
  • HTTP 404: report that the requested Gemini model is unavailable for the team. Do not silently choose a different model or provider.
  • Timeout, provider warning, or bad audio: stop and check Gateway usage before a user-authorized retry. Do not claim that an audio file was created or approved by listening.
  • Missing script or Node.js: report plugin setup failure. Do not improvise a token-bearing command.
  • Missing credentials: share the preset and script as a recipe, clearly labeled as not generated audio.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

Reviews can't be read right now.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free