1624318455/dsh-plugin-tts
Edge TTS voice plugin for DeepSeek Harness: read assistant replies aloud, auto-read toggle, voice settings panel (free, no API key)
dsh-plugin-tts is a dual-sided (Host + Web UI) DeepSeek Harness plugin that reads assistant replies aloud. Microsoft Edge's free TTS works out of the box with no API key. For custom voices, RVC voice models can be loaded from a voice-pack registry — a portable RVC runtime is included, so no separate RVC WebUI installation is needed. Long replies stream with gapless adaptive chunked playback. A read-aloud button appears on every finalized assistant message, and an auto-read toggle sits in the composer tool row.
Install
dsh plugin --profile web add "github:1624318455/dsh-plugin-tts#main"GitHub source install from main branch (no npm package; README-documented 2026-08-24). Dual-sided plugin (Host + Web UI). Edge TTS is free and works out of the box (no key). For RVC custom voices, download a portable RVC runtime from the voice-pack registry — no RVC WebUI install needed. See docs/USER-GUIDE.md for first-time setup and docs/RVC-GUIDE.md for custom voice configuration. Long replies use gapless adaptive chunked playback.
Compatibility
DeepSeek Harness web profile. Host + Web UI plugin (dual-sided). Microsoft Edge TTS (free, online, no key). Optional RVC custom voices via portable runtime. Node.js required on host side. 44 tests passing per CI.
Details
- Repo: 1624318455/dsh-plugin-tts
- Category: UI Enhancements
- Stars: 5
- Version: GitHub main branch (README-documented 2026-08-24; no npm package)
- Last push: 2026-08-17
- First seen: 2026-08-15
Recent updates
Current: Edge TTS free online TTS; RVC custom voice models from voice-pack registry; portable RVC runtime; gapless adaptive chunked playback; read-aloud button per message; auto-read toggle in composer; 44 smoke tests passing. See docs/RVC-GUIDE.md and docs/USER-GUIDE.md.
FAQ
- Does Edge TTS require an account or API key?
- No — Microsoft Edge TTS is free and works without any API key or account. The plugin accesses it as a standard online TTS service.
- How do I add a custom voice?
- Download a voice model from the voice-pack registry (one-click from within the plugin) and load it with the portable RVC runtime — no separate RVC WebUI installation is needed. See docs/RVC-GUIDE.md for step-by-step setup.
- How does gapless playback work for long replies?
- Long replies are split into adaptive chunks that are synthesized and queued for gapless sequential playback. Each chunk starts synthesis as early as possible so there is no silence gap between chunks. See docs/adaptive-chunked-playback.md for the design.
Alternatives
Alan2Z/dsh-speak · omdsh-dev/dsh-notification · omdsh-dev/dsh-session-health