mercury-voice

Mercury Voice Support

Mercury Voice is a voice client for Hermes Agent, an open-source coding agent that runs on your own computer. The app does not include or host the agent. You need a running Hermes server to use it.

Getting connected

1. Start your Hermes server. For use over the internet or a VPN, bind it behind TLS and configure a password (dashboard.basic_auth) or OAuth gate.
2. In Mercury Voice, enter the server address. You can paste the dashboard URL hermes prints at startup (the token is picked up automatically), an address like 192.168.1.5:9119, or a DNS name.
3. Tap Connect. If the server has a password gate, a sign-in form appears.
4. Pick a project and tap the microphone.

Common problems

“Couldn’t reach the server.” Check that the address and port are right and that your phone can reach the machine (same Wi-Fi, VPN, or a public DNS name with TLS). Plain HTTP works for IP addresses and .local names; DNS names default to HTTPS.

No spoken reply. The server needs a text-to-speech provider configured. The free edge provider works out of the box.

Nothing is transcribed. The server needs a speech-to-text provider with a valid key.

Microphone permission denied. Open Settings → Privacy & Security → Microphone and enable Mercury Voice.

Conversation stops when the screen locks. Make sure Background App Refresh and audio are allowed for the app, and that Low Power Mode is off.

Reporting a bug or asking a question

Open an issue: https://github.com/angel12/mercury-voice/issues

Include your iOS or macOS version, the Hermes version, and what you expected to happen.

Privacy

See the Privacy Policy. The app collects nothing and has no backend of its own.

This site is open source. Improve this page.