WebsiteCommunityDeploy your Lillo Agent

Overview

Overview

Lillo's chat system provides a robust, context-aware conversation engine powered by GPT-4o-mini. The system supports both private chats and group interactions, with intelligent message handling and response generation. It features multi-agent support, allowing different AI personalities to coexist and interact within the same chat environment.

Core Components

1. Message Handling

  • Message Reception: Processes incoming messages through webhook handlers

  • Response Triggers:

    • Direct mentions (@botname)

    • Private messages

    • Replies to bot messages

    • Command prefixes (/)

  • Multi-Agent Support:

    • Agent-specific configurations

    • Independent chat histories

    • Custom personalities

2. Context Management

  • Chat History: Maintains conversation context using PostgreSQL

  • Message Threading: Supports reply chains and conversation flow

  • State Persistence:

    • Tracks user and chat states

    • Maintains agent preferences

    • Caches recent interactions

3. Response Generation

  • AI Integration:

    • Uses GPT-4o-mini for natural language processing

    • Supports model preference per chat/agent

    • Dynamic system prompt generation

  • Function Calling: Supports tool integration for:

    • Image generation

    • Weather data

    • Market data (Gecko/Solana)

    • Time information

  • Message Formatting:

    • Handles mentions

    • Supports emojis

    • HTML formatting

    • Special character escaping

Database Schema

CREATE TABLE chat_messages (
  id SERIAL PRIMARY KEY,
  message_id BIGINT NOT NULL,
  chat_id BIGINT NOT NULL,
  user_id BIGINT NOT NULL,
  user_first_name VARCHAR(255) NOT NULL,
  username VARCHAR(255),
  role VARCHAR(50) NOT NULL,
  content TEXT NOT NULL,
  timestamp BIGINT NOT NULL,
  reply_to_message_id BIGINT,
  agent_id VARCHAR(255) NOT NULL,
  created_at TIMESTAMP WITH TIME ZONE DEFAULT CURRENT_TIMESTAMP,
  UNIQUE(chat_id, message_id, agent_id)
);

Message Flow

  1. Reception:

    User Message -> Webhook -> shouldRespondToMessage -> handleBotTag -> Response

  2. Processing:

    • Message cleaning and validation

    • Context retrieval (agent-specific)

    • System prompt generation

    • Model preference handling

    • Function call detection

  3. Response:

    • AI response generation

    • Function call handling

    • Message formatting

    • Reply sending

    • History update

Special Features

1. Welcome Messages

  • Automatic greeting for new group members

  • Personalized welcome messages

  • Context-aware introductions

  • Agent-specific greetings

2. Error Handling

  • Content policy violations

  • Rate limiting (per user/function)

  • Network errors

  • Invalid inputs

  • Database errors

  • User-friendly messages

3. Message Formatting

  • User mentions

  • Emoji support

  • HTML formatting

  • Reply threading

  • Special character escaping

4. Model Preferences

  • Per-chat model selection

  • Agent-specific defaults

  • Dynamic model switching

  • Preference persistence

Performance Optimization

  1. Database Indexing:

    CREATE INDEX idx_chat_messages_chat_id_timestamp 
ON chat_messages(chat_id, timestamp DESC);

CREATE INDEX idx_chat_messages_agent_id
ON chat_messages(agent_id);

  2. Caching:

    • Model preferences

    • Agent configurations

    • Recent chat history

    • Function call cooldowns

  3. Rate Limiting:

    • Function call cooldowns

    • API request throttling

    • Message rate limits

    • Per-user quotas

  4. Query Optimization:

    • Efficient history retrieval

    • Indexed lookups

    • Limited result sets

    • Chronological ordering

Security

  1. Input Validation:

    • Message content sanitization

    • Command parameter validation

    • User permission checks

    • Type safety enforcement

  2. Access Control:

    • Role-based permissions

    • Group membership validation

    • Command restrictions

    • Function call authorization

  3. Error Messages:

    • User-friendly error responses

    • Security-aware error details

    • Rate limit notifications

    • Content policy feedback

  4. Data Protection:

    • Secure token handling

    • Input sanitization

    • Output escaping

    • Error masking

Related Documentation

PreviousGrokNextAPI

Last updated