Chat History

Chat history enables multi-turn conversations. Without it, each run is isolated—the agent has no idea what was said before. With a database and add_history_to_context=True, previous messages are automatically included in every request.

Enable Chat History

Set add_history_to_context=True to include previous messages in every run:

1from kern.agent import Agent
2from kern.db.sqlite import SqliteDb
3
4agent = Agent(
5    db=SqliteDb(db_file="agent.db"),
6    add_history_to_context=True,
7    num_history_runs=3,  # Include last 3 turns
8)
9
10agent.print_response("My name is Sarah", session_id="chat_123")
11agent.print_response("What's my name?", session_id="chat_123")  # Agent knows: "Sarah"

1from kern.team import Team
2from kern.db.sqlite import SqliteDb
3
4team = Team(
5    db=SqliteDb(db_file="team.db"),
6    add_history_to_context=True,
7    num_history_runs=3,  # Include last 3 turns
8)
9
10team.print_response("My name is Sarah", session_id="chat_123")
11team.print_response("What's my name?", session_id="chat_123")  # Team knows: "Sarah"

Control History Size

More history means more tokens. Use these parameters to control what gets included:

1agent = Agent(
2    db=SqliteDb(db_file="agent.db"),
3    add_history_to_context=True,
4    num_history_runs=5,
5    num_history_messages=20,  # Cap at 20 messages total
6)

On-Demand History Access

Instead of always including history, let the agent decide when to look it up:

1agent = Agent(
2    db=SqliteDb(db_file="agent.db"),
3    read_chat_history=True,  # Agent gets a get_chat_history() tool
4)

The agent can call get_chat_history() when it needs context, rather than having history in every request. Useful for analytics, auditing, or when most queries don't need prior context.

Cross-Session History

Search across multiple sessions for context that spans conversations:

1agent = Agent(
2    db=SqliteDb(db_file="agent.db"),
3    search_session_history=True,
4    num_history_sessions=2,  # Search last 2 sessions
5)

Programmatic Access

Retrieve history directly in your code:

1# Get user-assistant message pairs
2chat_history = agent.get_chat_history(session_id="chat_123")
3
4# Get all messages from the session
5messages = agent.get_session_messages(session_id="chat_123")
6
7# Get the last run output with metrics
8last_run = agent.get_last_run_output()

Use this for building custom UIs, debugging, or exporting transcripts.

Team History

Teams support additional history sharing between members:

1team = Team(
2    db=SqliteDb(db_file="team.db"),
3    add_history_to_context=True,
4    num_history_runs=3,
5    add_team_history_to_members=True,  # Share history across team members
6)

With add_team_history_to_members=True, member agents see the full team conversation, not just their own interactions.

Workflow History

Workflows use add_workflow_history_to_steps to pass previous run results to steps:

1from kern.workflow import Workflow
2
3workflow = Workflow(
4    db=SqliteDb(db_file="workflow.db"),
5    add_workflow_history_to_steps=True,
6    num_history_runs=5,
7    steps=[...],
8)

Choosing a Pattern

Developer Resources

• Session Storage - What gets stored and how to retrieve it.
• AgentSession reference
• TeamSession reference