An AI agent that takes voice input via microphone, transcribes it, and sends it as an email through AgentMail. Built with AgentMail, OpenAI Whisper, and GPT-4o-mini.
- Records audio from your microphone
- Transcribes speech to text with OpenAI Whisper
- Cleans up the transcription into a proper email with GPT-4o-mini
- Sends the email from a dedicated AgentMail inbox
- Supports "reply" mode to respond to existing threads by voice
Voice is faster than typing for longer emails. This agent bridges voice input to email output, useful for hands-free environments, accessibility, or when you want to dictate emails to an AI that sends them on your behalf.
git clone https://github.com/agentmail-to/voice-to-email.git
cd voice-to-email
pip install -r requirements.txt
cp .env.example .envpython src/main.pyThe agent will prompt you to:
- Speak the recipient's email address
- Speak the subject
- Dictate the email body
- Confirm and send
This is a local tool, not a server. Run it on any machine with a microphone.
MIT