steve

hi, I'm steve. I'm an AI agent that ranks, scores and rewards contributions to a Discord server.

Prerequisites

Node.js 23.3.0

# Install nvm (Node Version Manager)
curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.39.3/install.sh | bash

# Install Node.js 23.3.0
nvm install 23.3.0

# Use Node.js 23.3.0
nvm use 23.3.0

pnpm

# Install pnpm
npm install -g pnpm

# Verify installation
pnpm --version

Postgres Database

1. Create a Supabase project (recommended)
2. Enable the pg_vector extension in the database settings of your supabase project

Discord Bot

1. Go to Discord Developer Portal

2. Create a new application

3. Go to the "Bot" section and create a bot

4. Enable these bot permissions:

   • Send Messages
   • Send Messages in Threads
   • Embed Links
   • Attach Files
   • Read Message History
   • Add Reactions
   • Connect (Voice)
   • Speak (Voice)
   • Use Voice Activity
   • Priority Speaker
   • Use Application Commands

5. Under "Privileged Gateway Intents", enable:

   • Message Content Intent
   • Server Members Intent
   • Presence Intent

6. Go to OAuth2 > URL Generator:

   • Select scopes: bot, applications.commands
   • Select the permissions listed above
   • Use the generated URL to invite the bot to your server

7. Save your bot token for later use

Local Development

1. Copy the .env.example file to .env and fill in the values

cp .env.example .env

2. Install dependencies

pnpm install

3. Run the bot

pnpm run dev

Environment Variables

Variable	Required	Description
DATABASE_URL	✅	Connection URL for your PostgreSQL database
POSTGRES_POOL_MIN	❌	Minimum number of connections in the database pool (recommended: 2)
POSTGRES_POOL_MAX	❌	Maximum number of connections in the database pool (recommended: 10)
OPENAI_API_KEY	⚡	OpenAI API key starting with 'sk-'. Required if using OpenAI as model provider
ANTHROPIC_API_KEY	⚡	Anthropic API key. Required if using Anthropic/Claude as model provider
DISCORD_APPLICATION_ID	✅	Your Discord application ID from the Developer Portal
DISCORD_API_TOKEN	✅	Your Discord bot token from the Developer Portal

✅ = Required
⚡ = Required (one model provider must be chosen)

Model Provider Selection

The model provider can be configured in src/characters/default.json:

{
  "modelProvider": "openai" // or "anthropic"
}

Scoring Rules

Steve uses a comprehensive scoring system defined in src/config/scoring-rules.json to evaluate and rank contributions. The scoring system consists of two main components:

Quality Score

Quality scoring evaluates the content of messages based on:

• Message Length (20% weight)
  • Minimum: 5 characters
  • Ideal: 100 characters
• Unique Words (20% weight)
  • Minimum: 3 words
  • Ideal: 20 words
• Relevance (30% weight)
  • Channel topic matching
  • Keyword matching
• Engagement (30% weight)
  • Reactions received
  • Replies generated

Trust Score

Trust scoring evaluates the contributor based on:

• Account Age (20% weight)
  • Minimum: 1 day
  • Ideal: 30 days
• Previous Contributions (30% weight)
  • Quality average
  • Quantity of contributions
• Community Standing (30% weight)
  • Server roles
  • Reputation score
• Report History (20% weight)
  • Rule violations
  • Warnings received

Minimum Requirements

For a message to be scored, it must meet these basic conditions:

• Minimum length: 1 character
• Minimum reactions: 2
• Minimum quality score: 0
• Minimum trust score: 0

You can customize these scoring rules by modifying the src/config/scoring-rules.json file. The schema for this configuration is defined in src/config/scoring-rules.schema.ts.

Scoring Actions

When a message meets the scoring conditions, you can configure custom actions to be executed. These actions are defined in TypeScript files under src/clients/discord/actions/.

Here's an example of a scoring action that sends a cat fact when conditions are met:

// src/clients/discord/actions/scoring_action.ts
import type { Message, TextChannel } from "discord.js";
import type { ScoringAction } from "../types";
import { sendMessageInChunks } from "../utils";

export const scoringAction: ScoringAction = {
  name: "cat-facts",
  handler: async (message: Message, meetsConditions: boolean) => {
    if (!meetsConditions) return;

    const catFactsResponse = await fetch("https://catfact.ninja/fact");
    const catFactsData = await catFactsResponse.json();
    const catFact = catFactsData.fact;

    await sendMessageInChunks(
      message.channel as TextChannel,
      `You met the scoring conditions! Here's a cat fact: ${catFact}`,
      message.id,
      []
    );
  },
};

To create your own scoring action:

1. Create a new file in src/clients/discord/actions/
2. Define your action using the ScoringAction interface:

   interface ScoringAction {
     name: string;
     handler: (message: Message, meetsConditions: boolean) => Promise<void>;
   }

3. Export your action as the default export
4. The handler will be called automatically when a message meets the scoring conditions

Your action can do anything you want - send messages, add reactions, update a database, or integrate with external services.

Character Customization

You can customize the AI character by modifying src/characters/default.json. The character file supports:

• name: Character's display name
• bio: Array of biographical statements
• lore: Array of backstory elements
• topics: Subjects the character is knowledgeable about
• style: Behavior patterns for different contexts
• messageExamples: Sample conversations
• postExamples: Sample social media posts
• adjectives: Character traits

For detailed documentation on character configuration, see the Character Files documentation.

Deployment

Using Fly.io (Recommended)

1. Install the Fly CLI:

# macOS
brew install flyctl

# Windows
powershell -Command "iwr https://fly.io/install.ps1 -useb | iex"

# Linux
curl -L https://fly.io/install.sh | sh

2. Sign up and log in:

fly auth signup
# or if you already have an account
fly auth login

3. Launch your application:

fly launch

This will:

• Create a fly.toml file
• Create and configure a Fly.io app

4. Set your environment variables:

fly secrets set DISCORD_API_TOKEN="your-token"
fly secrets set DISCORD_APPLICATION_ID="your-app-id"
fly secrets set OPENAI_API_KEY="your-key"
fly secrets set DATABASE_URL="your-postgres-url"
# Set any other required environment variables

5. Deploy your application:

fly deploy

Your bot should now be running in production! You can monitor its status with:

fly status
fly logs