Agent Platform — Hermes Agent Voice Chat with Memory & Deep Research
Agent Platform is the voice-first home for Hermes Agent, an installable AI assistant. Hold the on-screen mic button (or hold the spacebar on desktop) to talk; Hermes Agent transcribes your speech, routes the request to Claude, GPT-5, or Gemini, and can speak its reply back via OpenAI or Microsoft Azure neural voices.
Why Agent Platform
- Built around Hermes Agent. Persistent memory keeps facts and preferences across every conversation on Agent Platform.
- Hold-to-talk anywhere. Push-to-talk voice input with live amplitude feedback. No browser TTS — top-tier speech models only.
- Deep research, with citations. Hermes Agent plans, runs multi-step web searches across distinct angles, and cites primary sources.
- Multi-model. Route Hermes Agent across Anthropic Claude Opus 4.7 / Sonnet 4.6 / Haiku 4.5, OpenAI GPT-5 / GPT-5 Mini, Google Gemini 3 Pro / Flash.
- Installable PWA. Add Agent Platform to your home screen; offline app shell, mobile-first layout.
How Hermes Agent works on Agent Platform
Tap the microphone to start, release to send. The audio is transcribed by a top-tier speech-to-text model (OpenAI gpt-4o-mini-transcribe or Azure Speech). Hermes Agent streams a reply through your chosen language model, and Agent Platform can speak the response automatically with OpenAI gpt-4o-mini-tts or Azure Neural TTS. Every chat is saved as a transcript.
Frequently asked questions
What is Agent Platform?
Agent Platform is a voice-first installable web app that lets you talk to Hermes Agent with a hold-to-talk button, persistent memory, and deep research across the web.
What is Hermes Agent?
Hermes Agent is the conversational AI served by Agent Platform. It supports voice input, spoken replies, deep research, and routes to top-tier models including Claude, GPT-5, and Gemini.
Which models can Hermes Agent use?
Hermes Agent on Agent Platform supports Anthropic Claude Opus 4.7, Claude Sonnet 4.6, Claude Haiku 4.5, OpenAI GPT-5 / GPT-5 Mini, and Google Gemini 3 Pro / Gemini 3 Flash. Voice runs on OpenAI or Microsoft Azure Speech.
JavaScript is required to use the live Agent Platform application.