TTS Wyoming Piper Plugin: plugin for Hermes Agent

Hermes Agent text-to-speech plugin that uses a remote Piper server over the Wyoming protocol

What TTS Wyoming Piper Plugin does

Wyoming is a peer-to-peer TCP protocol for voice assistants from the Open Home Foundation, the team behind Home Assistant and Rhasspy. Piper is a local neural text-to-speech engine that speaks it, runs on CPU and needs no API key. This plugin connects Hermes Agent directly to a Piper instance, for example on a Raspberry Pi or home server, without Home Assistant in between. The README lists 44 languages and 174+ pre-trained voices.

Piper produces raw PCM, which the plugin converts with ffmpeg to the chosen format: MP3 by default, OGG for voice bubbles on Telegram, Matrix, WhatsApp and Signal, WAV, or FLAC. A voice_compatible setting tells Hermes whether the output is already voice-bubble ready so it can skip its own conversion. Two modes exist: pipe, which sends PCM straight into ffmpeg in one pass, and stream, which supports streaming delivery. The plugin warms up and reuses connections and ships a benchmark script. Host, port, voice, timeout, mode and format are set in config.yaml.

Key features

Connects to any Wyoming-compatible Piper server
44 languages and 174+ pre-trained voices, selectable by config or per request
Output as MP3 (default), OGG, WAV or FLAC
Pipe and stream synthesis modes
Warm-up and connection reuse for repeated synthesis
Benchmark script for profiling formats and modes on your hardware

When to use it

Giving a Hermes agent offline speech output from Piper on a Raspberry Pi
Sending voice-bubble replies to Telegram or Matrix through OGG output
Sharing one Piper server between Home Assistant and Hermes

Who it is for: Home lab users and privacy-focused Hermes users who run Piper locally and want speech output without cloud APIs.

How it fits with Hermes Agent

Built as a Hermes Agent TTS plugin. It is selected with tts.provider set to wyoming-piper in ~/.hermes/config.yaml.

How to install TTS Wyoming Piper Plugin

These commands are copied from the project's README. Check the repository for the latest steps before you run them.

hermes plugins install scross01/hermes-tts-wyoming-piper-plugin --enable

Requirements: A Wyoming-compatible Piper server reachable from Hermes (default host localhost, port 10200) and ffmpeg for format conversion

FAQ

What is TTS Wyoming Piper Plugin?

It is a Hermes Agent plugin that generates speech by calling a remote Piper server over the Wyoming protocol. It converts Piper's raw audio to MP3, OGG, WAV or FLAC.

How do I install TTS Wyoming Piper Plugin?

Run hermes plugins install scross01/hermes-tts-wyoming-piper-plugin --enable, then restart the desktop app or gateway. Set tts.provider to wyoming-piper and add host, port and voice under plugins.entries.tts-wyoming-piper.settings.

Which output format should I use for Telegram?

For Telegram or Matrix only, use output_format ogg with voice_compatible true. If you use both the TUI and Telegram, keep mp3 with voice_compatible false and let Hermes convert to OGG when needed.