Goa-AI

Transform Your Agent Development

Goa-AI extends Goa’s design-first workflow into a complete framework for building agentic, tool-driven systems in Go. Describe your agents, toolsets, policies, and workflows once in Goa’s DSL, then let Goa-AI generate everything you need: typed tool payloads, Temporal workflows, runtime registries, and durable execution loops.

What Makes Goa-AI Different?

Goa-AI stands out by treating your agent design as a living contract, just like Goa does for APIs:

• Type-safe tools: Generated payload/result structs plus JSON codecs—no more hand-written schemas
• Durable orchestration: Temporal-ready workflows & activities with retry/timeouts baked in
• Agent composition: Treat one agent as a tool of another, even across processes
• Operational visibility: Stream planner/tool/assistant events; persist transcripts; instrument with logs/metrics/traces
• MCP integration: Consume tool suites from MCP servers through generated wrappers

How Goa-AI Works

DSL → Codegen → Runtime → Engine + Features

From a single design file, Goa-AI generates:

1. Agent Packages - Workflow definitions, planner activities, and registration helpers
2. Tool Specs - Typed payload/result structs, JSON codecs, and JSON Schema definitions
3. Runtime Integration - Durable plan/execute loop with policy enforcement, memory persistence, and streaming hooks
4. MCP Adapters - Generated wrappers that bridge MCP servers into the runtime

The result is a cohesive architecture where planners focus on business logic while Goa-AI supplies the plumbing for Temporal, Mongo-backed memory, Pulse streams, MCP integration, and model providers.

Next Steps

• Follow the Getting Started guide to build your first agent
• Explore the Core Concepts to understand the DSL, runtime, and toolsets
• Check out the Tutorials for step-by-step examples
• Learn about Real-World Patterns for production deployments