| name | hibiki-tts |
| description | Text-to-speech via OpenAI API with optional summarization and translation using the hibiki CLI. |
Hibiki TTS Skill
Use the hibiki CLI to speak text aloud using OpenAI's text-to-speech API, with optional AI-powered summarization and translation.
When to Use
Invoke this skill when the user:
- Wants text read aloud
- Asks to hear something spoken
- Needs to summarize and speak content
- Wants text translated and spoken
- Requests audio output of text content
- Wants a file read aloud
Requirements
- The Hibiki macOS app must be running (check menu bar for speaker icon)
- The CLI communicates with the app via URL scheme
Commands
Basic Text-to-Speech
hibiki --text "Hello, world!"
Speaks the provided text aloud using the configured voice.
File Input (Text from File)
hibiki --file-name README.md
Reads UTF-8 text from a file and speaks it aloud.
Summarize + TTS
hibiki --text "Long article or document text here..." --summarize
hibiki --file-name notes.md --summarize
Summarizes the text using an LLM, then speaks the summary aloud. Useful for long content.
Summarize + TTS (Custom Prompt)
hibiki --text "Long article..." --summarize --prompt "Concise: 3 bullets, no fluff."
hibiki --text "Design doc..." --summarize --prompt "Thorough: goals, decisions, tradeoffs, risks, next steps."
Overrides the default summarization prompt (requires --summarize).
Translate + TTS
hibiki --text "Hello" --translate ja
hibiki --file-name draft.txt --translate fr
Translates the text to the specified language, then speaks it aloud.
Full Pipeline (Summarize + Translate + TTS)
hibiki --text "Long article text..." --summarize --translate fr
Summarizes the text, translates the summary, then speaks it aloud.
Help
hibiki --help
Supported Languages
| Code | Language |
|---|
en | English |
ja | Japanese |
de | German |
fr | French |
es | Spanish |
Example Usage
User asks: "Read this paragraph aloud"
hibiki --text "The paragraph content here"
User asks: "Summarize this article and read it to me"
hibiki --text "Full article content..." --summarize
User asks: "Give me a concise summary and read it"
hibiki --text "Full article content..." --summarize --prompt "Provide a concise summary: 2 bullets."
User asks: "Give me a thorough summary and read it"
hibiki --text "Full article content..." --summarize --prompt "Provide a thorough summary: key points, evidence, caveats, next steps."
User asks: "Translate this to Japanese and speak it"
hibiki --text "Hello, how are you?" --translate ja
User asks: "Summarize this and read it in German"
hibiki --text "Long content..." --summarize --translate de
User asks: "Read this file out loud"
hibiki --file-name docs/notes.txt
Notes
- Maximum text length is ~32KB (after URL encoding)
- Input source: use exactly one of
--text or --file-name
--file-name expects UTF-8 text files- Audio plays through the Hibiki AudioPlayerPanel UI
- Requests are saved to Hibiki's history
- Voice and API settings are configured in the Hibiki app's Settings
- The command runs asynchronously - audio plays after a brief processing delay
- For very long content, use
--summarize to condense before speaking
