Doubao TTS — Raycast Extension

Select any text on macOS, read it aloud via Raycast — powered by Volcengine Doubao TTS V3.

---

Why Doubao TTS?

Doubao TTS is the leading Chinese AI speech synthesis engine. It delivers unmatched naturalness, emotional expression, and voice diversity for Chinese text. Whether you're listening to research papers, long articles, or everyday text, Doubao TTS provides near-human-quality speech — no extra apps required.

Who is this for?

• Researchers — Listen to papers and documents, free your eyes
• Developers — Review docs, READMEs, and comments by ear
• Language learners — Hear standard Chinese pronunciation
• Content creators — Preview how your text sounds
• Anyone who needs TTS — Select text, press a key, listen

Features

• Quick Read — Select text, read aloud instantly (no UI)
• Voice Selection — Browse 90+ voices organized by category
• Stop Reading — Stop playback anytime
• Toggle mode — Trigger Quick Read again to stop
• Smart chunking — Auto-split long text by sentence
• Model switching — TTS 2.0 (recommended) and TTS 1.0
• Chinese & English — Built-in voices for both languages

Screenshots

Installation

Prerequisites

• Raycast installed
• A Volcengine account with Doubao TTS enabled (guide below)

Steps

1. Search "Doubao TTS" in the Raycast Store, install the extension
2. Configure your App ID and Access Key on first launch (guide below)
3. Bind a hotkey to Quick Read Selected Text (how?)
4. Select any text, press the hotkey — done!

Configuration

Raycast will prompt for preferences on first use.

Setting	Description	Required
App ID	Volcengine app identifier	✅
Access Key	Volcengine access key	✅
Model Version	TTS model (default: 2.0)
Default Voice	Voice for Quick Read
Speech Rate	Playback speed (0.5x–2.0x)

Get App ID & Access Key

1. Sign up and log in to Volcengine Console
2. Go to Speech → Doubao TTS
3. Enable the service if not already active
4. Find your credentials:
   • App ID = X-Api-App-Id
   • Access Key (Access Token) = X-Api-Access-Key
5. See also: Console FAQ

Tip: New Volcengine users get a free quota. Check the console for details.

Model Versions

Model	Resource ID	Description
Doubao TTS 2.0 (Recommended)	seed-tts-2.0	Latest model, best quality
Doubao TTS 1.0	seed-tts-1.0	Classic model, more voices
Doubao TTS 1.0 (High Concurrency)	seed-tts-1.0-concurr	Higher concurrency
Voice Clone 2.0	seed-icl-2.0	Voice cloning
Voice Clone 1.0	seed-icl-1.0	Voice cloning

Note: Different models support different voices. TTS 2.0 shows only 2.0 voices; TTS 1.0 shows only 1.0 voices.

Voice List

90+ built-in voices organized by category:

Category	Examples	Model
General Female	Vivi, Xiaohe, Cancan	1.0 / 2.0
General Male	Yunzhou, Xiaotian, Qingcang	1.0 / 2.0
Emotional Female	Emotional Cancan, Sweet Female	1.0
Emotional Male	Emotional Male	1.0
English	Tim, Adam, Amanda	1.0 / 2.0
Japanese / Korean / Multilingual	Japanese Female, Korean Female	1.0 / 2.0
Fun Accents / Role Play	Dongbei Bro, Beijing Accent	1.0

Full voice list: Doubao Voice Catalog

Usage

Quick Read (Recommended)

1. Select text in any app
2. Open Raycast (⌥ Space)
3. Type Quick Read and press Enter
4. It reads aloud! Trigger again to stop

Bind a Hotkey (Highly Recommended)

Bind a global hotkey to Quick Read for the ultimate workflow: select text → press hotkey → instant reading, no need to open Raycast every time.

1. Open Raycast → search Extensions
2. Find Doubao TTS
3. Click Record Hotkey next to Quick Read Selected Text
4. Press your desired key combo (e.g. ⌥ R, ⌃ ⌥ S)
5. Done! Select text anywhere and press the hotkey to read

Tip: You can also bind a hotkey to Stop Reading for quick stopping.

Read with Voice Selection

1. Select text
2. Open Read with Voice Selection in Raycast
3. Browse voices, pick one
4. Press Enter to start

Stop Reading

• Run Stop Reading in Raycast
• Or trigger Quick Read again while playing

Development

Project Structure

raycast-doubao-tts/
├── src/
│   ├── api/
│   │   ├── volcengine-tts.ts   # V3 API client
│   │   └── types.ts            # TypeScript types
│   ├── constants/
│   │   └── voices.ts           # 90+ voice configs
│   ├── utils/
│   │   ├── audio-player.ts     # Audio player (afplay)
│   │   └── text-chunker.ts     # Smart text chunking
│   ├── quick-read.tsx          # Quick Read command
│   ├── read-with-voice.tsx     # Voice selection command
│   └── stop-reading.tsx        # Stop playback command
├── assets/
│   └── command-icon.png        # Extension icon
├── package.json
└── tsconfig.json

Local Development

npm install    # Install dependencies
npm run dev    # Dev mode (hot reload)
npm run build  # Build
npm run lint   # Lint

Technical Details

• API: Volcengine Doubao TTS V3 HTTP unidirectional streaming
• Auth: HTTP Headers (X-Api-App-Id, X-Api-Access-Key, X-Api-Resource-Id)
• Response: JSON Lines (NDJSON), one JSON object per line
• Audio: MP3, 24000 Hz
• Chunking: Smart split by punctuation, ≤1024 UTF-8 bytes per chunk
• Playback: macOS built-in afplay
• Cross-command stop: PID file ($TMPDIR/doubao-tts.pid)

References

• Raycast Extension Docs
• Doubao TTS V3 HTTP API
• Doubao Voice Catalog
• Volcengine Console FAQ

Acknowledgements

• Bob Plugin - Doubao TTS — Inspired the configuration approach
• Volcengine — Doubao TTS API provider

License

MIT

---

中文文档: README.zh.md