The History Tools provide access to Discord chat history for context-aware responses.
HistoryTools
Toolkit for reading and analyzing chat history.
class HistoryTools(Toolkit):
def __init__(self)
read_chat_history
Read chat history for the current channel.
async def read_chat_history(self, limit: int = 2000) -> str
Maximum number of messages to retrieve
Formatted chat history as a string, or error message
Behavior
- Gets channel from execution context
- Fetches recent messages from cache and database
- Falls back to database-only if channel object unavailable
- Returns formatted history with timestamps and authors
Each message is formatted as:
2024-03-04 10:30:00 Username(user_id): Message content
Example Usage
from tools.history_tools import HistoryTools
from core.execution_context import set_current_channel
# Setup
history_tools = HistoryTools()
set_current_channel(message.channel)
# Read history
history = await history_tools.read_chat_history(limit=100)
print(history)
In Agent Context
When used with agents, the tool is automatically available:
from agno.agent import Agent
from tools.history_tools import HistoryTools
agent = Agent(
name="Context Agent",
tools=[HistoryTools()],
instructions="""Use read_chat_history to analyze conversations.
Always fetch at least 5000 messages on first try."""
)
# Agent can now call the tool
result = await agent.arun(
"What did John say about the project yesterday?"
)
Error Handling
Returns descriptive error messages:
- No context:
"Error: No execution context found. Cannot determine channel."
- No history:
"No history found in database."
- Fetch error:
"Error fetching history: <error details>"
- Uses
get_recent_context() which combines cache + database
- Automatically fetches from Discord API if cache insufficient
- Limits prevent excessive memory usage
- Async operations don’t block other tasks
Data Sources
The tool retrieves messages from:
- In-memory cache (fastest, recent messages)
- PostgreSQL database (persistent storage)
- Discord API (fetches missing messages)
Recommended Limits
- Quick context: 100-200 messages
- Conversation analysis: 1000-2000 messages
- Deep history search: 5000+ messages
- Maximum: 50000 (configurable via
CONTEXT_AGENT_MAX_MESSAGES)
Integration Example
Complete example with agent team:
from agno.agent import Agent
from agno.team import Team
from tools.history_tools import HistoryTools
from core.execution_context import set_current_channel
# Create context QnA agent
context_agent = Agent(
name="Chat Context Q&A",
role="Answering questions about users, topics, and past conversations",
model=model,
tools=[HistoryTools()],
instructions="""
You specialize in answering questions about chat history.
IMPORTANT: Always fetch a minimum of 5000 messages on first try.
Use the history to:
- Answer "who said what" questions
- Summarize discussions on specific topics
- Track when topics were last mentioned
- Identify user opinions and statements
- Provide context about past conversations
Be precise with timestamps and attribute statements accurately.
"""
)
# Use in team
team = Team(
name="AI Team",
members=[context_agent],
model=model
)
# In message handler
@bot.event
async def on_message(message):
set_current_channel(message.channel)
response = await team.arun(
"What did we discuss about AI last week?",
session_id=str(message.channel.id)
)
await message.channel.send(response.content)
Configuration
Related environment variables:
CONTEXT_AGENT_MAX_MESSAGES: Maximum messages for context (default: 50000)POSTGRES_URL: Database connection for persistent history
Dependencies
core.database.get_messages: Database query functioncore.execution_context: Channel context trackingdiscord_bot.context_cache.get_recent_context: Cache retrieval