Deploy a ZeroClaw Agent on ClawMate

Name your agent and select ZeroClaw

In the deploy wizard, enter a name and select "ZeroClaw" as the framework. ZeroClaw uses a Rust runtime with just 256MB RAM, starts in under 10ms, and costs 10x less to run than OpenClaw. Ideal for simple, focused agents that need speed and low overhead.

Choose your AI model and paste your API key

ZeroClaw supports 22+ model providers. Select Anthropic (Claude), OpenAI (GPT-5), Groq (Llama, free tier), or OpenRouter (300+ models). Paste your API key — encrypted with AES-256-GCM. Configure fallback models for automatic retry on rate limits.

Connect Telegram or Discord

Select your messaging platform and enter your bot token. Configure DM policy (open or restricted to specific user IDs). Enable optional voice transcription with Groq Whisper for voice messages.

Set personality and deploy

ZeroClaw uses TOML-based configuration for simplicity. Choose a system prompt template or write your own. Configure security settings — exec sandbox enabled by default, network access disabled by default. Click Deploy and your agent goes live in seconds.

# ZeroClaw deployment timeline:
# 1. Container created
# 2. Rust binary starts (<10ms)
# 3. Channels connected
# 4. Agent is live
# Total: ~15-30 seconds