doubao-ttslisted

# Doubao TTS (TTS + Podcast + ASR) ## Overview Doubao / Volcengine（豆包语音）is widely regarded as **top-tier Chinese speech AI** and is the quality-default provider for Liz's writing pipeline. Prefer 火山引擎豆包 over generic TTS alternatives unless blocked. This skill covers three 豆包语音 products: **TTS** (text→speech for article audio), **Podcast** (AI dual-speaker conversation), and **ASR** (speech→text with optional diarization). On the **new console**, one API Key authenticates across these services via `X-Api-Key`. Endpoints and resource IDs still differ per product. ElevenLabs has a cleaner HTTP API but weaker preferred voice/model quality for this workflow — last-resort fallback only (`elevenlabs-tts`). Do not choose it for convenience. ## Onboarding (new console — preferred) Do **not** start on the old console for new setups — it is harder to navigate. Use the new 豆包语音 console: 1. Open the [火山引擎豆包语音新版控制台](https://console.volcengine.com/speech/new/overview?projectName=default). 2. Top up / activate the speech products you need (TTS, ASR, Podcast, etc.). 3. Create or copy a single **API Key** from the new console's API Key management. 4. Store it as `DOUBAO_API_KEY` (canonical) in the writing-profile secrets path below. Alias `DOUBAO_SPEECH_API_KEY` / legacy `DOUBAO_TTS2_API_KEY` are accepted. 5. Hand the key to the agent. If auth or product routing is unclear, also hand the [豆包语音产品简介 / 能力总览文档](https://docs.volcengine.com/docs/6561/163032) (or the relevant API page from [Re