Getting Started

Install

go install github.com/DocumentDrivenDX/agent/cmd/ddx-agent@latest

Quick Start with LM Studio

Start LM Studio and load a model with tool-calling support (e.g., Qwen 3.5).
Run ddx-agent:

ddx-agent -p "Read main.go and tell me the package name"

DDX Agent connects to LM Studio at localhost:1234 by default.

Quick Start with Anthropic

export AGENT_PROVIDER=anthropic
export AGENT_API_KEY=sk-ant-...
export AGENT_MODEL=claude-sonnet-4-20250514

ddx-agent -p "Read main.go and tell me the package name"

Configuration

Create .agent/config.yaml in your project:

provider: openai-compat
base_url: http://localhost:1234/v1
model: qwen3.5-7b
max_iterations: 20
session_log_dir: .agent/sessions

Environment variables override the config file:

AGENT_PROVIDER — openai-compat or anthropic
AGENT_BASE_URL — provider base URL
AGENT_API_KEY — API key
AGENT_MODEL — model name

As a Library

import (
    "context"
    "github.com/DocumentDrivenDX/agent"
    "github.com/DocumentDrivenDX/agent/provider/openai"
    "github.com/DocumentDrivenDX/agent/tool"
)

func main() {
    p := openai.New(openai.Config{
        BaseURL: "http://localhost:1234/v1",
        Model:   "qwen3.5-7b",
    })

    result, err := agent.Run(context.Background(), agent.Request{
        Prompt:   "Read main.go and tell me the package name",
        Provider: p,
        Tools: []agent.Tool{
            &tool.ReadTool{WorkDir: "."},
            &tool.BashTool{WorkDir: "."},
        },
        MaxIterations: 10,
    })
    // result.Output contains the agent's response
}

Session Replay

Every run is logged. Replay past sessions:

ddx-agent log                  # list sessions
ddx-agent replay <session-id>  # human-readable replay