AI Agent Hub
Back to plugins
🧰

dsh-plugin-elevenlabs-callback

Web Tools Updated 2026.09.02

Run the following command in DeepSeek Harness:

dsh plugin install StephenEvenson/dsh-plugin-elevenlabs-callback

Paste the following prompt into your AI chat to install this plugin:

Run dsh plugin install StephenEvenson/dsh-plugin-elevenlabs-callback in your terminal to install this plugin from https://github.com/StephenEvenson/dsh-plugin-elevenlabs-callback.

About this plugin

Hand a refactor to DeepSeek Harness and walk away. Twenty minutes later the turn finishes, or it stalls on an rm command waiting for approval, and a link appears in the terminal that you will never see. dsh-plugin-elevenlabs-callback closes that attention gap: when a run ends, fails, or needs approval, it pushes a link to your phone. Tapping it opens a conversation with an ElevenLabs voice agent that has already read the full run log, summarises the outcome in a sentence, and asks what to do next.

The core capability is a complete voice round-trip. Once you confirm the wording, the agent calls one of three webhook tools the plugin provisions on your ElevenLabs workspace (get_run_details, send_instruction, resolve_approval) and the instruction lands back in the dsh session as a follow-up turn. In approval scenarios the keyboard prompt and the voice call race in parallel; whichever answers first wins and the other is told it was already handled.

Who is it for? You run long tasks (cross-file refactors, batch scripts, multi-turn debugging sessions) and want to be notified and able to dictate the next step without sitting at the terminal. If you prefer ntfy pushes or desktop notifications, you can use the link-only path and skip the phone call entirely. With a Twilio number you can have the agent ring your handset directly. The ElevenLabs free tier (about 15 agent-minutes per month) covers light daily use.

Use Cases

  • A multi-file refactor finishes while you are away from the desk and a voice call reports the outcome on your phone
  • The harness stalls on an rm command awaiting approval and the voice agent calls you to ask whether to grant it
  • A run fails and you want a spoken summary before deciding to retry, without going back to the terminal

Best For

  • Developers who run long autonomous tasks and want mobile awareness plus remote instruction without sitting at the terminal
  • Engineers already using ElevenLabs voice agents who want to wire them into their AI coding workflow
  • Remote workers who manage async workflows via Twilio ringing calls or ntfy push notifications