Description

This plugin adds espeak-ng as a text-to-speech engine for OpenVoiceOS. espeak-ng runs fully offline and supports many languages and voices. The plugin registers with ovos-plugin-manager so any OVOS component can select it as a TTS backend.

Install

pip install ovos_tts_plugin_espeakng

The plugin needs the espeak-ng executable. Install it with:

apt-get install espeak-ng

Configuration

Add this block to your OVOS configuration to select the plugin and a voice:

  "tts": {
    "module": "ovos_tts_plugin_espeakng",
    "ovos_tts_plugin_espeakng": {
      "voice": "m1"
    }
 }

Docker

This plugin also ships a self-contained image that serves espeak-ng behind ovos-tts-server's ElevenLabs-compatible HTTP API on port 9666. espeak-ng runs fully offline, so the container needs no network access and no API keys at runtime.

Pull the published image:

docker run --rm -p 9666:9666 ghcr.io/openvoiceos/ovos-tts-plugin-espeakng:dev

Or build it locally. You can override the default language at build time:

docker build -t ovos-tts-plugin-espeakng --build-arg LANG=pt .
docker run --rm -p 9666:9666 ovos-tts-plugin-espeakng

Or use the bundled compose file:

docker compose up

Synthesize speech with a request to the running server:

curl 'http://localhost:9666/synthesize/hello%20world' --output hello.wav

Related projects

• OpenVoiceOS/ovos-plugin-manager: loads and manages OVOS TTS plugins.
• OpenVoiceOS/ovos-tts-server: HTTP server that exposes this and other TTS plugins over an API.
• espeak-ng/espeak-ng: the speech synthesizer this plugin wraps.