| 12345678910111213141516 |
- name: hermes-live-voice
- repo: https://github.com/Synero/hermes-live-voice
- sha: 0578e0cc35022a9dcb469cb9fbfc4d8afc93fc65
- version: "0.2.6"
- description: "Full-duplex Live Voice for Hermes Desktop — GPT-Live-1 on the ChatGPT/Codex subscription via the local Codex app-server, with live transcript, voice tool calls and task delegation to the user chat. Disclosure — subscription voice uses your Hermes openai-codex sign-in through Hermes-signed requests (requires_auth; the plugin never holds or refreshes the token) to mint OpenAI Realtime client secrets, and the usage panel reads Hermes's own account-usage fetcher; the gpt-live-1 lane runs the local codex app-server on the Codex CLI's own login, and the dashboard can run codex login --device-auth, copying ~/.codex/auth.json (or $CODEX_HOME) to a backup before that login, restoring the backup if the login is cancelled, and reading the file read-only for sign-in status; with TALK_OPENAI_API_KEY or OPENAI_API_KEY set it uses that key instead; if hermes-talk is installed it loads that plugin's modules from ~/.hermes/plugins/hermes-talk and substitutes its own talk_auth; voice audio goes to OpenAI over WebRTC, and spoken tasks are submitted as messages in your focused chat."
- maintainer: Synero
- tier: community
- category: voice
- requires_hermes: ">=0.21.7"
- docs_url: https://github.com/Synero/hermes-live-voice
- platforms: []
- capabilities:
- provides_tools: []
- provides_hooks: []
- provides_middleware: []
- requires_env: []
|