Typing is fine, but sometimes you just want to talk. OpenClaw supports voice interaction through multiple methods — from Telegram voice messages to full talk mode with wake words.
Voice Options Overview
| Method | Setup Difficulty | Cost | Best For |
|---|---|---|---|
| Telegram voice messages | None | Free | Quick voice notes |
| Talk Mode (browser) | Easy | API costs for TTS/STT | Hands-free chat |
| Phone calls | Medium | Telephony provider fees | On-the-go access |
| iOS/Android app | Hard | Free (local) | Mobile voice assistant |
Method 1: Telegram Voice Messages (Easiest)
The simplest voice interaction requires zero setup. Just send a voice message to your OpenClaw bot on Telegram.
How it works:
- Open your bot in Telegram
- Hold the microphone button and speak
- OpenClaw receives the audio, transcribes it, and responds with text
Pros: Works immediately, no extra configuration Cons: Responses are text-only (no voice reply)
This works with any OpenClaw setup, including ClawTank.
Method 2: Talk Mode (Browser-Based)
Talk Mode turns OpenClaw into a conversational voice assistant in your browser — like having a video call with your AI.
Components:
- STT (Speech-to-Text): Converts your voice to text (Whisper)
- TTS (Text-to-Speech): Converts AI responses to voice (ElevenLabs, OpenAI TTS)
- Continuous loop: Listen → Think → Speak → Listen
Setup:
- Enable TTS in OpenClaw config:
openclaw config set tts.provider "elevenlabs"
openclaw config set tts.apiKey "YOUR_ELEVENLABS_KEY"
- Enable STT:
openclaw config set stt.provider "openai"
- Access Talk Mode through the OpenClaw Control UI in your browser.
Pros: Full voice conversation, hands-free Cons: Requires browser, additional API costs
Method 3: Phone Calls
Yes, you can call your OpenClaw assistant on an actual phone number.
How it works: OpenClaw integrates with telephony providers (Plivo, Twilio, Telnyx) through the Voice Call plugin. You dial a number, and your AI assistant picks up.
Setup overview:
- Sign up for a telephony provider (Twilio is most popular)
- Get a phone number
- Install the voice call plugin:
openclaw plugins install voice-call
- Configure with your provider credentials
- Call your number and start talking
Pros: Access from any phone, anywhere Cons: Telephony costs, more complex setup, slight latency
Method 4: Mobile Voice Assistant
iOS
The iOS app supports voice chat natively. After pairing with your Gateway, you can have voice conversations directly from your iPhone.
Android
The OpenClaw Assistant app for Android features:
- Customizable wake words
- Long-press home button activation
- Continuous conversation mode
- Offline wake word detection
Pros: Always-available voice assistant on your phone Cons: Requires building from source, Gateway must be running
Choosing a TTS Voice
The voice your assistant uses matters. Options:
- ElevenLabs — Most natural-sounding, many voice options, $5/mo+
- OpenAI TTS — Good quality, simple setup, pay-per-use
- Local (Piper) — Free, runs on your server, less natural but zero latency
For most users, ElevenLabs offers the best experience.
Voice on ClawTank
Currently, ClawTank supports text interaction through Telegram, including Telegram voice messages (which get transcribed automatically).
Full Talk Mode and phone call features are on the roadmap. In the meantime, Telegram voice messages provide a solid voice-to-text experience out of the box.
The Practical Recommendation
Start with Telegram voice messages — it's free, requires no setup, and works today. If you find yourself wanting voice responses, add ElevenLabs TTS to your OpenClaw configuration.
Full voice assistant mode (wake words, continuous conversation) is powerful but complex to set up. Get comfortable with the text-based experience first.
