Hermes Speech

Hermes Speech is a plugin that connects Hermes Agent to AllModels, a voice AI API service. With it installed, Hermes can understand voice messages, reply out loud, find voices and speech models, and manage an AllModels balance through normal conversation. Setup is done by sending a message to the Hermes agent that points to the plugin's GitHub repository, then restarting the Hermes gateway. Voice works in several places. In the CLI and terminal UI, a user enters /voice on, optionally /voice tts, presses Ctrl+B and speaks, and Hermes detects the pause, sends the text and answers aloud. The desktop app has a start voice conversation button. In Telegram, Hermes transcribes a voice note and replies with a playable voice message, and the documentation also covers Discord voice channels. It is for people who already run Hermes Agent and want spoken input and output. AllModels itself offers speech to text, text to speech, voice cloning and realtime streaming, with SDK compatibility for OpenAI and ElevenLabs style clients. On pricing, AllModels states that a free tier carries a 0% platform fee with one free model and one free provider, and that pay as you go carries a 5.5% platform fee over the downstream provider's cost, with 73 models across 18 providers. The vendor says it does not mark up provider pricing, taxes may apply when an account is topped up, and credits may be expired after 12 months.