hermes-whatsapp-voice plugin
A Hermes Agent plugin for live WhatsApp voice calls: you call the bot or it calls you, with speech handled by Gemini Live.
- 155 ★ on the repository
- Last commit
- MIT
Broad access
Why this level
- The plugin drives calls through a WhatsApp linked device and handles the owner voice.
- Audio goes to Gemini Live, and on the free tier Google may use the data.
- Calls cost money on the paid tier.
Install
- Works in
Manual install
hermes plugins install ashwanthbalakrishnan5/hermes-whatsapp-voice --enableThe manual install command from the README; it needs Hermes 0.21.6 or newer and a Gemini key first.
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- The plugin drives calls through a WhatsApp linked device and handles the owner voice.
- Audio goes to Gemini Live, and on the free tier Google may use the data.
- Calls cost money on the paid tier.
This is third-party code. Review the repository files before installing.
What it does
The plugin adds a WhatsApp voice channel to Hermes. You can call the bot number and hand the agent a task mid-conversation, such as moving a meeting or finding a message. For urgent cases Hermes calls you: a reminder, an alert, a cron job that needs a decision. If you do not answer, it falls back to another channel. Speech runs on Gemini Live in speech-to-speech mode. The voice takes its name, personality and memory from your Hermes. A Rust sidecar downloads on first run, and number pairing and call history live in the WhatsApp Voice tab of the Hermes dashboard. After a call a short summary arrives in WhatsApp, and the transcript with timings is available through `hermes whatsapp-voice calls`.
Who it is for. Hermes Agent users who want to talk to it by voice over WhatsApp and get calls from it about important events.
Good fit when
- The agent should ring you when a chat message is easy to miss.
- Tasks are easier to delegate by voice on the move than by typing.
- A cron job waits for the owner decision and cannot sit until morning.
- You want voice access to the agent without a WhatsApp Business or Cloud API account.
Not a fit when
- Several people need to call: the bot answers only the owner number.
- Voice cannot be sent to the external Gemini service.
- You need an official WhatsApp Business channel with a contract and support.
Example request
Call me if the CI build fails and ask whether to roll the release back.Limitations
Needs Hermes 0.21.6 or newer on Linux or macOS, a separate WhatsApp number for the bot and a Gemini key in Hermes. Google login (Cloud Code OAuth) does not work for the Live API. On the free tier Google may use the data; on the paid tier a call costs roughly 3 to 5 cents a minute by the author estimate. The link works as a linked device rather than the Business API, so check WhatsApp terms for that setup yourself.
How to disable. Disable the plugin in Hermes and unlink the device in WhatsApp (linked devices). Delete the Gemini key and stored call records separately.
FAQ
Who can call the bot?
Only the owner. The plugin answers one configured number and calls only that number, and no tool accepts a destination, so prompt injection cannot make it call anyone else.
Is WhatsApp Business required?
No. A personal number linked as a device works. The bot needs its own number, and in `self-chat` mode you cannot call yourself.
How do I link the bot number?
Open the WhatsApp Voice tab in the Hermes dashboard, click Pair with WhatsApp and scan the QR code from the bot phone. In a terminal use `hermes whatsapp-voice pair`.
Where can I see call history?
Run `hermes whatsapp-voice calls`: every call keeps a transcript and timings.
How much does a call cost?
On the paid Gemini tier roughly 3 to 5 cents a minute by the author estimate. A free tier exists, but Google may use the data.
Discussion
No questions or reviews yet.
Related
PluginsAlice Voice for Hermes
hermes-alice-voice
Voice input through Yandex Alice for Hermes Agent: a private skill sends recognized text to a personal Telegram dialog. Experimental, for a single operator
A Hermes Agent plugin with a Relay server: chat, voice and a terminal on Android, plus a desktop CLI for the agent's tools
A self-improving agent from Nous Research with a TUI, messaging gateway, cron jobs and skills it writes itself
Official Hermes Agent plugins: Home Assistant, Spotify, Blender, Snyk, Langfuse, sandboxes and importers for other agents
A pack of 19 plugins for Hermes Desktop: provider quotas, an iteration meter, a memory review queue, session management and RSS.
A Herdr plugin that gives every coding agent its own Telegram topic, with status in the icon and messages both ways.
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Repository stars | Main difference |
|---|---|---|---|
| hermes-whatsapp-voice (this card) | no | 155 | |
| WhatsApp AgentKit | no | 556 | scaffolds a text WhatsApp bot; the plugin adds two-way voice calls to Hermes |
| Hermes Tool Router | no | 145 | routes Hermes tools; the plugin adds a voice channel |
README in short
Features, requirements, one-line agent install or a manual command, number pairing, language setup, CI call tests and security.