Convert agent events into TTS audio routed over SSH on Tailscale devices.
npx skills add https://github.com/SouthpawIN/burner-phone --skill speak-southpawin
Please help me install this Agent Skill. Skill: speak Source: https://github.com/SouthpawIN/burner-phone/tree/main/skills/speak Command: npx skills add https://github.com/SouthpawIN/burner-phone --skill speak-southpawin
The speak Skill solves the problem of silent automation by narrating task acknowledgments, reasoning checkpoints, tool/file actions, and final responses through text-to-speech across your Tailscale devices.
Enable narration and have the agent speak a message on the default Duo device by running: "speak --on" then "speak "Starting the build now." --if-on --device duo".
High-intent search queries and answers about installing and using this skill.
To add TTS narration to an automation agent, you route agent events like task acknowledgments, tool actions, and final responses to generate text-to-speech audio. This provides continuous audio feedback for silent build and automation loops.
Yes, you can send text-to-speech audio over SSH to multiple devices on Tailscale. The system auto-detects reachable devices like a Surface Duo 2 or local speakers and routes spoken playback directly to them across the network.
Queued audio playback for agent narration works by handling speech asynchronously in the background. When audio is already playing, additional messages are queued to preserve chronological ordering without blocking the agent's ongoing execution loop.
Yes, enabling agent narration over SSH requires an operational TTS backend like Soprano via HTTP. You also need device SSH reachability through Tailscale to successfully route and play the generated speech.
You gate agent speech using the optional --if-on flag to ensure narration only occurs when speak mode is enabled. This prevents unwanted audio output by checking the mode status before routing text-to-speech messages to devices.
The best way to narrate agent reasoning and tool operations is by integrating calls at deterministic points in the agent loop: acknowledge, analyze, act, progress, and complete. This captures comprehensive text-to-speech updates across your devices.
plurigridCoordinates ssh, curl, nmap and Tailscale for remote Linux administration and diagnostics.
SvtooSynthesize local offline speech via the Piper TTS engine.
AcosmiAutomate sub-agent triggering with cron, event, and message_match triggers.
TitoPrauseeDetect Mercury peer changes in Tailscale meshes via port 9443 probes.
