Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
Version History
**Expanded configuration options, improved security clarity, and optional ElevenLabs support** - Added support for optional ElevenLabs API key for voice responses. - Updated documentation to clarify required vs. optional credentials and environment variables. - Improved security section: clearly states capabilities and limitations. - Expanded configuration instructions for Deepgram and new ElevenLabs option. - Added best practices and validation/testing steps for easier setup and safer usage.
What This Skill Does
Sends emails from natural voice commands received through Telegram. Audio is transcribed by Deepgram and parsed into recipient, subject, and body fields, then sent through Gmail via gogcli. Built for accessibility, it requires no typing.
It lets users send emails entirely by voice through an existing Telegram conversation, with no additional app or interface required.
When to use it
- Dictating emails while driving
- Hands-free email for users with limited mobility
- Sending quick replies from a Telegram voice message
- Composing emails without access to a keyboard
- Voice-only email workflow for low-vision users
Example Workflow
Here's how your AI assistant might use this skill in practice.
User sends Telegram voice message: "new email to mom@gmail.com, subject Dinner, body See you at 7pm, send"
- 1Receives voice message via Telegram
- 2Transcribes audio using Deepgram nova-3 model
- 3Parses recipient, subject, and body from transcription
- 4Sends email via gogcli using authenticated Gmail account
- 5Returns a text or voice confirmation to the user
Email sent to mom@gmail.com with subject "Dinner" and body "See you at 7pm", confirmation returned in chat
Requirements
Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.