Skip to main content

​
Overview

The SmartResponse Addon for Drako Bot is a modular, AI-driven response system designed for Discord servers. It supports multiple AI providers, handles multi-step interactions, and integrates seamlessly with internal knowledge bases for context-aware responses. The system is built with extensibility in mind, allowing developers to add new AI providers or customize bot behavior without modifying core logic.

​
Features

  • Automatic Node.js module management Missing modules are detected and installed on first run.
  • Multi-provider AI support via Vercel AI SDK Supports OpenAI, Groq, Mistral, and easily extendable to other providers.
  • Multi-step response handling with interactive Discord buttons.
  • Internal knowledge base integration for enhanced context-aware replies.
  • Advanced message matching using semantic similarity and stemmed matching.
  • Discord modal and button interactions for user-friendly input.
  • Conversation caching with automatic timeout management.
  • Modular and maintainable architecture for easier future extensions.
Note: Auto-installation occurs for missing modules, ensuring smooth first-time setup.

​
Configuration

All bot settings are defined in config.yml.

​
SmartResponses

Configure default AI behavior and internal data sources: 
SystemPrompt: "You are a helpful Discord assistant."
InternalDataSources:
  - "resources/docs"
  - "resources/wiki"
  - "resources/knowledge_base"
Triggers:
  WhitelistedChannels: ["channel-id-1", "channel-id-2"]
  WhitelistedCategories: ["category-id-1"]
  RequireMention: true

​
AI Providers

Define API keys, models, and response settings: 
AIProvider: 'openai'

Providers:
  OpenAI:
    APIKey: 'your-openai-api-key'
    Model: 'gpt-4'
    Temperature: 0.7
    MaxTokens: 1024
  Groq:
    APIKey: 'your-groq-api-key'
    Model: 'llama3-70b-8192'
    Temperature: 0.7
    MaxTokens: 1024
  Mistral:
    APIKey: 'your-mistral-api-key'
    Model: 'mistral-large-latest'
    Temperature: 0.7
    MaxTokens: 1024

​
Phrases

Define triggers and responses: 
Phrases:
  Greeting:
    Phrase: ["hello", "hi", "hey"]
    Type: TEXT
    Response: "Hello! How can I assist you today?"
    MatchPercent: 0.8

​
Adding New AI Providers

The modular architecture makes it straightforward to add new AI providers. The following guide outlines the required steps.

​
Step 1: Install Provider SDK

Install the official SDK for your AI provider: 
npm install @ai-sdk/<provider-name>

​
Step 2: Create Provider Module

Create a new file in providers/, e.g., anthropic.js: 
const { createAnthropic } = require('@ai-sdk/anthropic');
const { generateText } = require('ai');

async function generateResponse(smartResponse, message, phrase) {
  const providerConfig = smartResponse.config.Providers.Anthropic;
  const input = {
    model: providerConfig.Model,
    prompt: phrase || message.content,
    max_tokens: providerConfig.MaxTokens,
    temperature: providerConfig.Temperature
  };
  const response = await generateText(createAnthropic(providerConfig.APIKey), input);
  return response.text;
}

module.exports = { generateResponse };

​
Step 3: Update Providers Index

Include the new provider in providers/index.js: 
const anthropic = require('./anthropic');

​
Step 4: Update SmartResponse Logic

Modify SmartResponse.generateAIResponse to handle the new provider: 
switch(this.config.AIProvider.toLowerCase()) {
  case 'openai':
    return openai.generateResponse(this, message, phrase);
  case 'groq':
    return groq.generateResponse(this, message, phrase);
  case 'mistral':
    return mistral.generateResponse(this, message, phrase);
  case 'anthropic':
    return anthropic.generateResponse(this, message, phrase);
  default:
    throw new Error(`AI provider ${this.config.AIProvider} not supported`);
}

​
Step 5: Configure Provider in YAML

Add the new provider configuration in config.yml: 
Anthropic:
  APIKey: 'your-anthropic-api-key'
  Model: 'claude-3-opus-20240229'
  Temperature: 0.7
  MaxTokens: 1024

​
Step 6: Test Your Provider

  • Restart the bot and confirm the new provider works.
  • Send a test message in a whitelisted channel.
  • Ensure the AI responds as expected.

​
Usage

​
Bot Interaction

  • The bot listens to messages in whitelisted channels and categories.
  • If RequireMention is enabled, mention the bot to trigger a response.
  • Multi-step responses use Discord buttons:
    • Show Next Step: Proceed to the next message in a sequence.
    • Ask Another Question: Opens a modal for new input.

​
Internal Knowledge Base

  • Place your reference documents in resources/docs/, resources/wiki/, or resources/knowledge_base/.
  • The bot automatically indexes and retrieves context from these sources.

​
Conversation Cache

  • Responses are cached to improve performance.
  • Cache timeout ensures memory is not consumed indefinitely.

​
Automatic Module Management

The bot verifies required modules on runtime:
  • Logs missing dependencies.
  • Installs them automatically via npm install.
  • Continues execution after installation.
Auto-installed modules: 
natural, fs, path, js-yaml, discord.js,
@ai-sdk/openai, @ai-sdk/groq, @ai-sdk/mistral, @ai-sdk/google, @ai-sdk/togetherai, ai

​
Credits

  • John – Owner of Drako Bot and original creator of SmartResponse.
⌘I