Production-ready Go framework for building intelligent multi-agent AI systems

AgenticGoKit enables developers to build sophisticated agent workflows with dynamic tool integration, multi-provider LLM support, and enterprise-grade orchestration patterns. Designed for Go developers who need the performance and reliability of compiled binaries with the flexibility of modern AI agent systems.

⚠️ Alpha Release: AgenticGoKit is currently in alpha development. APIs may change, and some features are still being developed. Suitable for experimentation and early adoption, but not recommended for production use yet.

# Install CLI
go install github.com/kunalkushwaha/agenticgokit/cmd/agentcli@latest

# Create your first agent project
agentcli create my-agent --template basic
cd my-agent

# Set up environment (Azure OpenAI example)
export AZURE_OPENAI_API_KEY=your-key
export AZURE_OPENAI_ENDPOINT=https://your-resource.openai.azure.com/
export AZURE_OPENAI_DEPLOYMENT=your-deployment

# Run it
go run .

package main

import (
    "context"
    "github.com/kunalkushwaha/agenticgokit/core"
)

func main() {
    // Create agents
    agents := map[string]core.AgentHandler{
        "researcher": NewResearchAgent(),
        "analyzer":   NewAnalysisAgent(),
    }
    
    // Run collaborative workflow
    runner := core.CreateCollaborativeRunner(agents, 30*time.Second)
    result, err := runner.ProcessEvent(ctx, event)
}

• Multi-Agent Orchestration: Parallel, sequential, and loop patterns
• Memory & RAG: PostgreSQL pgvector, Weaviate, in-memory providers
• Tool Integration: MCP protocol for dynamic tool discovery
• CLI Scaffolding: Generate complete projects instantly
• Configuration: TOML-based with environment overrides
• Visualization: Auto-generated Mermaid workflow diagrams

my-agent/
├── main.go              # Entry point
├── agentflow.toml       # Configuration
├── go.mod               # Go module
├── agents/              # Agent implementations
└── docs/                # Generated diagrams

# agentflow.toml
[orchestration]
mode = "collaborative"
timeout_seconds = 30

[agent_memory]
provider = "pgvector"
enable_rag = true

[mcp]
enabled = true

# Available project templates
agentcli create my-project --template basic           # Single agent
agentcli create my-project --template research        # Research team
agentcli create my-project --template rag-system      # Knowledge base
agentcli create my-project --template chat-system     # Conversational

import (
    _ "github.com/kunalkushwaha/agenticgokit/plugins/llm/ollama"
    _ "github.com/kunalkushwaha/agenticgokit/plugins/memory/pgvector"
    _ "github.com/kunalkushwaha/agenticgokit/plugins/orchestrator/default"
)

Check out examples/ for working demos:

• Simple Agent
• Multi-Agent Collaboration
• RAG Knowledge Base
• MCP Tool Integration

# Clone and build
git clone https://github.com/kunalkushwaha/agenticgokit.git
cd agenticgokit
make build

# Run tests
make test

# Run examples
cd examples/01-simple-agent
go run .

• Website: www.agenticgokit.com
• Documentation: docs.agenticgokit.com
• Examples: examples/
• Discussions: GitHub Discussions
• Issues: GitHub Issues

We welcome contributions! See docs/contributors/ContributorGuide.md for getting started.

Apache 2.0 - see LICENSE for details.