AgentField
ActiveDescription
A framework for building, running, and scaling AI agents as APIs and microservices, with built-in observability, auditability, and identity-aware access control from day one.
Key Features
- Build AI agents as REST endpoints with Python, Go, or TypeScript — no DSL or YAML required
- Cryptographic identity and tamper-proof audit trails for every agent execution
- Built-in human-in-the-loop with pause/resume for approval workflows via webhooks
- Multi-agent orchestration with routing, coordination, and cross-agent call tracing
- Prompt-to-production: describe a system in one line and get a Docker Compose stack
- Harness Orchestration for multi-turn coding agents with Claude Code, Codex, Gemini CLI, and OpenCode
Use Cases
💡 Deploying production AI agents as microservices with built-in observability and auditability
💡 Building insurance claims processors or other approval-required workflows with human escalation
💡 Creating multi-agent backends where agents call each other through a control plane
💡 Running AI agents with versioned deployments, canary releases, and A/B testing
💡 Integrating AI agents into existing service stacks as callable API endpoints
Categories
Quick Start
Install with `curl -fsSL https://agentfield.ai/install.sh | bash`, then scaffold an agent with `af init my-agent --defaults`. Start the control plane with `af server` and run your agent with `python main.py`. Call it via `curl -X POST http://localhost:8080/api/v1/execute/my-agent.demo_echo`.