Python client and CLI for Volcengine/ByteDance Doubao seed-tts-2.0 bidirectional streaming TTS.
If you already use Doubao, the doubao mcp server is the piece that lets your assistant work with it directly. Python client and CLI for Volcengine/ByteDance Doubao seed-tts-2.0 bidirectional streaming TTS.
doubao-speech is the first PyPI package that covers both directions of Volcengine's modern voice stack:
The toolset is worth reading before you wire it up, because it tells you what the integration is really for:
Punctuation — Automatic commas, periods, question marksLow — latency streaming**: 200ms chunks by default; tune via --segment-msTTS — One synthesize call opens a fresh WebSocket and tears it down atSTT — Streaming starts returning partial transcripts within ~500 msv0.2 — connection-reuse daemon (TCP+TLS amortization), streamingv0.3 — LangChain/LlamaIndex/Dify integration recipesv1.0 — API frozen, semver guaranteesendpoint — Behaviourbigmodel_async — Bidirectional, optimized; emits only when text changesbigmodel_nostream — Streaming-input; returns after >15 s or the final packetDirection — Input / OutputAlias — LanguageConfiguration is passed through the environment: HERMES_LOCAL_STT_COMMAND, VOLCENGINE_APP_ID, VOLCENGINE_ACCESS_TOKEN, DOUBAO_APP_ID, DOUBAO_ACCESS_TOKEN. Treat anything key-shaped as a real credential — scope it to the minimum the server needs, and rotate it if it ever lands in a shared config.
The server ships on PyPI as doubao-speech, so your MCP client can launch it on demand — there is no separate build step. Add the server block to your client's configuration, restart it, and the tools register themselves.
This sits in the AI and media services group, where several servers overlap in what they claim to do but differ sharply once you actually set them up. Doubao's toolset — Punctuation, Low, TTS and 11 more — is a fair guide to whether it matches your workflow.
This entry was verified against Doubao's own documentation before publication; SyncDev keeps the directory reviewed rather than auto-generated.
| Tool | What it does |
|---|---|
| Punctuation | Automatic commas, periods, question marks |
| Low | latency streaming**: 200ms chunks by default; tune via --segment-ms |
| TTS | One synthesize call opens a fresh WebSocket and tears it down at |
| STT | Streaming starts returning partial transcripts within **~500 ms**. |
| v0.2 | connection-reuse daemon (TCP+TLS amortization), streaming |
| v0.3 | LangChain/LlamaIndex/Dify integration recipes. |
| v1.0 | API frozen, semver guarantees. |
| endpoint | Behaviour |
| bigmodel_async | Bidirectional, optimized; emits only when text changes |
| bigmodel_nostream | Streaming-input; returns after >15 s or the final packet |
| Direction | Input / Output |
| Alias | Language |
| zh-female-reporter | zh-CN |
| zh-male-warm | zh-CN |
{
"mcpServers": {
"doubao-tts": {
"command": "uvx",
"args": ["doubao-speech"],
"env": {
"HERMES_LOCAL_STT_COMMAND": "your-value",
"VOLCENGINE_APP_ID": "your-value",
"VOLCENGINE_ACCESS_TOKEN": "your-value",
"DOUBAO_APP_ID": "your-value",
"DOUBAO_ACCESS_TOKEN": "your-value"
}
}
}
}Add to claude_desktop_config.json, then restart Claude Desktop.
| Variable | Description | Required |
|---|---|---|
| HERMES_LOCAL_STT_COMMAND | Configuration value read at startup. | Optional |
| VOLCENGINE_APP_ID | Configuration value read at startup. | Optional |
| VOLCENGINE_ACCESS_TOKEN | Credential the server authenticates with. | Yes |
| DOUBAO_APP_ID | Configuration value read at startup. | Optional |
| DOUBAO_ACCESS_TOKEN | Credential the server authenticates with. | Yes |
Build a programmable telecommunications stack for connecting telephony services with the Internet via a cloud-based utility.
Search built for AI, not humans — semantic web search that returns model-ready content, plus code context.
Answers, not links — delegate questions to Perplexity's search-grounded models and get cited responses back.
Give your assistant a voice — text-to-speech, voice cloning and audio tools from the ElevenLabs API.
Give your assistant a real code sandbox — isolated cloud VMs for actually running the code it writes.
The ML hub in your context window — search models, datasets, papers and run Spaces from the official server.