Open Source iOS Client

Talk to your AI agents.

Voice and text chat with your self-hosted OpenClaw agents. On-device speech recognition. Streaming responses. Multi-agent channels.

$2.99 one-time purchase — or build from source for free

ClawTalk icon App screenshot coming soon
Features

Everything you need to talk to your agents.

Built for voice-first interaction with privacy at its core.

Voice-First Design

Push-to-talk for quick messages or hands-free conversation mode with Voice Activity Detection. Interrupt anytime.

On-Device Speech Recognition

WhisperKit runs entirely on your iPhone's Neural Engine. Your audio never leaves the device — not even to your own server.

Streaming Responses

Text streams in real-time with full markdown rendering — code blocks, tables, lists, and inline formatting.

Three TTS Options

Hear responses spoken aloud. Choose ElevenLabs for quality, OpenAI for value, or Apple's built-in voice for free offline TTS.

Multi-Agent Channels

Create separate channels for different agents. Each channel maintains its own conversation history and agent configuration.

Image Attachments

Attach up to 8 photos per message for visual context. Your agent sees the images and responds to what's in them.

Tools Dashboard

Browse agent memory, view sessions, check browser status, and select models — all without going through chat.

Security First

Keychain-encrypted credentials, HTTPS-only connections, on-device STT, and encrypted chat history. No analytics, no tracking.

Open Source

MIT licensed. Build it yourself for free from GitHub, or grab the App Store version for convenience and automatic updates.

How It Works

From voice to response in seconds.

A direct connection between your phone and your agents. No intermediary servers.

You speak

Push-to-talk or hands-free. Audio stays on your device.

WhisperKit transcribes

On-device speech recognition via Apple Neural Engine.

Your agent responds

Streamed over HTTPS to your self-hosted OpenClaw gateway.

ClawTalk speaks

Response spoken aloud while text streams with full markdown.

Your iPhone
WhisperKit (STT)
Text-to-Speech
Chat UI + Markdown
←——→ HTTPS / Tailscale / Cloudflare Tunnel
Your Server
OpenClaw Gateway
AI Agents
Tools + Memory
Security

Your data stays yours.

No analytics. No tracking. No telemetry. Everything stays on your device or your server.

Keychain Storage

All credentials encrypted by the Secure Enclave.

HTTPS Enforced

Plain HTTP connections are rejected. TLS 1.2 minimum.

On-Device STT

Audio is transcribed locally. Never uploaded anywhere.

Encrypted History

Chat history protected by iOS Data Protection at rest.

No Open Ports

Connect via Cloudflare Tunnel or Tailscale. No exposed services.

No Telemetry

Zero analytics, zero tracking, zero data collection. Period.

One-Time Purchase
$2.99

No subscription. No in-app purchases. Yours forever.

  • All features included
  • Automatic updates via App Store
  • Open source — MIT license
  • Build from source for free
  • Support ongoing development
Download on the App Store

or build from source on GitHub

Requires iOS 17.0+ and a self-hosted OpenClaw gateway