Overview

The quickest way to get started is using Docker Compose, which includes Reactive Agents, PostgREST, and PostgreSQL in a containerized environment.

Prerequisites

Step 1: Clone and Setup

# Clone the repository
git clone https://github.com/idkhub-com/reactive-agents.git
cd reactive-agents

Step 2: Start with Docker Compose

docker-compose up

docker-compose up -d

docker-compose logs -f

Step 3: Access Dashboard

Once Docker Compose has started all services, open your browser and visit:

http://localhost:3000

Configuration Steps

Initial Dashboard

To enable optimization features, complete these steps:

1. Add an AI Provider

Navigate to the AI Providers & Models section and click "Add AI Provider." Configure your AI provider settings and click "Create AI Provider."

2. Add a Model

After adding an AI provider, you are automatically redirected to the Models setup section, were you will add your first AI model from that provider. Configure your first model and click "Add 1 Model".

Successful AI Provider and Model looks like:

3. Create an Agent

Navigate to the Agents tab.

Click "Create Agent"

Configure your first agent to start handling AI interactions, and click "Create Agent"

4. Create Skills

Add skills to define what your agent can do. Click "Create Skill"

Configure the skill details including name, description, partitions and advanced parameters.

5. Add Models and Evaluations to Skills

Associate models with your skills by selecting the appropriate model from the available options. Select from available evaluations to configure evaluation criteria for your skill to measure and optimize agent performance. Finally, Click "Complete Setup"

Successful Skill Setup looks like:

6. Test Your Setup

Make your first API call to verify everything is working. Replace agent_name and skill_name with the names you created in the previous steps:

import OpenAI from 'openai';
import 'dotenv/config';

const client = new OpenAI({
  apiKey: process.env.BEARER_TOKEN,
  baseURL: 'http://localhost:3000/v1',
});

const idkhubConfig = {
  targets: [
    {
      optimization: 'auto'
    },
  ],
  agent_name: 'My Assistant',
  skill_name: 'General Help',
};

const response = await client
  .withOptions({
    defaultHeaders: {
      'ra-config': JSON.stringify(idkhubConfig),
    },
  })
  .chat.completions.create({
    model: 'gpt-5',
    messages: [
      { role: 'user', content: 'Hello, how are you?' }
    ],
  });

console.log(response.choices[0].message.content);

const idkhubConfig = {
  targets: [
    {
      optimization: 'auto',
    },
  ],
  agent_name: 'My Assistant',
  skill_name: 'General Help',
};

const response = await fetch('http://localhost:3000/v1/chat/completions', {
  method: 'POST',
  headers: {
    'Authorization': `Bearer ${process.env.BEARER_TOKEN}`,
    'Content-Type': 'application/json',
    'ra-config': JSON.stringify(idkhubConfig),
  },
  body: JSON.stringify({
    model: 'gpt-5',
    messages: [
      { role: 'user', content: 'Hello, how are you?' }
    ],
  }),
});

const data = await response.json();
console.log(data.choices[0].message.content);

#!/bin/bash
curl -X POST http://localhost:3000/v1/chat/completions \
  -H "Authorization: Bearer ${BEARER_TOKEN}" \
  -H "Content-Type: application/json" \
  -H "ra-config: {\"targets\":[{\"optimization\":\"auto\"}],\"agent_name\":\"My Assistant\",\"skill_name\":\"General Help\"}" \
  -d '{
    "model": "gpt-5",
    "messages": [
      {"role": "user", "content": "Hello, how are you?"}
    ]
  }'

Next Steps

• Explore the Dashboard: Familiarize yourself with the interface
• Configure Multiple Providers: Add support for different AI providers
• Set Up Evaluations: Create custom evaluation criteria
• Monitor Performance: Track your AI system's performance over time

Ready to dive deeper? Check out our Core Concepts guide.