songjongdevsecops/dsh-plugin-voz-copilot

Gives the DeepSeek Harness push-to-talk session a name and a sidebar home, adds composer dictation, and can hold a hands-free spoken conversation.

Names the push-to-talk session and gives it a place in the sidebar, so voice conversations stop being anonymous entries, and adds dictation microphones to the composer. Its real-time mode keeps the microphone open while a Silero VAD decides when you are speaking, starts transcription while you are still silent, and speaks the answer back sentence by sentence while the model is still writing. The plugin is the glue between the harness and a speech stack running on localhost, removing the package removes its bundle line, and conversations that were already named keep their title in the session log.

Web UI Enhancements ★ 0

Source repository unavailable. The repository songjongdevsecops/dsh-plugin-voz-copilot no longer exists on GitHub (HTTP 404 when checked on 2026-10-28), so the install command and any further commands on this page cannot currently be run. They are kept below as a record of what the plugin documented while it was available.

Repository unavailable ↗

Install

npx @deepseek-ai/dsh plugin --profile web add github:songjongdevsecops/dsh-plugin-voz-copilot --allow-build=onnxruntime-node

Copied verbatim from the README's Install block. --allow-build=onnxruntime-node is mandatory on pnpm 10+ — the README states that without it pnpm blocks the ONNX Runtime binary download and the install fails, and that the package itself has no build step. No npm package is published, so the GitHub spec is the install path. Restart the harness afterwards; the bundle layer is read at boot.

Compatibility

Node.js ^22.19.0 or >=24.0.0, pnpm on PATH, and the web profile for the interface (the title mode is used by the acp profile). The plugin ships no speech-to-text engine and no text-to-speech voice: it expects an HTTP endpoint compatible with whisper.cpp's whisper-server at whisperUrl, default http://127.0.0.1:8178, answering POST /inference?response_format=json with the audio in the multipart file field, and that server must decode WebM/Opus (an ffmpeg-enabled build), WAV and MP3. An optional faster engine can be set with whisperFastUrl.

Details

Alternatives

mervyn-teo/dsh-plugin-qr-connect · omdsh-dev/dsh-plugin-skills · ZG2017/dsh-plugin-model-icons

More plugins in Web UI Enhancements

Browse more in Web UI Enhancements

Guides for Web UI Enhancements plugins