Chat Handler

​

async_ask_junkie

async def async_ask_junkie(
    user_text: str,
    user_id: str,
    session_id: str,
    images: list = None,
    client=None
) -> str

​

Example

from discord_bot.chat_handler import async_ask_junkie
from agno.media import Image

images = [Image(url="https://example.com/image.png")]
response = await async_ask_junkie(
    user_text="What's in this image?",
    user_id="123456789",
    session_id="channel_123",
    images=images
)

​

setup_chat

def setup_chat(bot)

​

Registered Events

​

on_ready

• Initializes database connection
• Connects MCP tools
• Starts message backfill task
• Syncs offline edits/deletes

​

on_message

• Handles chatbot prefix (!) messages
• Builds context from recent chat history
• Processes images from messages and replies
• Sets execution context for tools
• Generates and sends AI responses

​

on_message_edit

• Updates message cache when messages are edited

​

on_message_delete

• Updates message cache when messages are deleted

​

on_disconnect

• Closes database connections gracefully

​

Example

from discord_bot.selfbot import SelfBot
from discord_bot.chat_handler import setup_chat

bot = SelfBot(token="...", prefix="!")
setup_chat(bot)
bot.run()

​

main_cli

async def main_cli()

​

Behavior

• Creates a team for default CLI user
• Runs interactive CLI if terminal is available
• Gracefully closes MCP tools on exit
• Skips CLI app in non-interactive environments

​

Example

import asyncio
from discord_bot.chat_handler import main_cli

asyncio.run(main_cli())

​

Message Processing Flow

1. Message received → Cache updated
2. Prefix detected (!) → Extract prompt
3. Mentions resolved → Convert to readable format
4. Context built → Include recent messages (default: 100)
5. Reply context → Include replied message if present
6. Images extracted → From message and reply
7. Execution context set → Channel ID and object
8. AI team invoked → Generate response
9. Mentions restored → Convert back to Discord format
10. Response sent → With timing information

​

Context Building

• Retrieving recent messages from cache (limit: TEAM_LEADER_CONTEXT_LIMIT)
• Including replied message for conversation threading
• Resolving user mentions to readable names
• Attaching image URLs from messages and replies
• Formatting with timestamps and user information

​

Error Handling

• Invalid responses → Returns fallback error message
• Team failures → Logs error and sends truncated error to user
• Backfill failures → Logs error but doesn’t block bot startup
• MCP connection failures → Logs warning and continues