Agent Zero is a next-generation autonomous AI assistant framework that runs in its own virtual computer environment. It features dynamic tool creation, persistent memory, and multi-agent cooperation capabilities.
- Autonomous AI agent with code execution capabilities
- Runs in isolated Docker container with full Linux environment
- Web-based UI for interaction
- Multi-agent cooperation system
- Persistent memory across sessions
- Dynamic tool creation and adaptation
- No predefined limitations
- Deploy the SDL file using Akash Console or CLI
- Wait for deployment to complete
- Access the web UI through the provided URI on port 80
- Configure your AI model provider in Settings
- OpenAI (recommended for best results)
- Anthropic Claude
- Local models via Ollama
- Set authentication credentials in Settings for security
You must configure an AI model provider before Agent Zero will function:
- Go to Settings → Agent Settings
- Select your provider (OpenAI, Anthropic, or Ollama)
- Enter your API key or model configuration
- Save settings
- Enable authentication: Settings → Authentication
- Set UI login username and password
- Set root password for SSH access if needed
- CPU: 1.0 units
- Memory: 2Gi
- Storage: 10Gi
- Port: 80 (HTTP web interface)
- Initial container startup takes approximately 30-60 seconds
- Configure AI provider immediately after deployment
- Enable authentication if exposing to internet
- Data persists within container storage allocation
- Uses approximately 2GB for Docker image
- Official Website: https://agent-zero.ai
- GitHub Repository: https://github.com/agent0ai/agent-zero
- Docker Hub: https://hub.docker.com/r/agent0ai/agent-zero
- Documentation: https://agent-zero.ai/p/docs/get-started/
Current deployment version: v0.9.8
For Agent Zero specific issues, visit the GitHub repository or official Discord. For Akash deployment issues, visit Akash Network Discord.