Learning Machines

Agents that learn and improve with every interaction.

Kern turns agents into learning machines by combining them with Learning Stores. Learning Stores are persistent backends that capture user profiles, memories, and knowledge over time.

1from kern.agent import Agent
2from kern.db.sqlite import SqliteDb
3from kern.models.openai import OpenAIResponses
4
5agent = Agent(
6    model=OpenAIResponses(id="gpt-5.2"),
7    db=SqliteDb(db_file="tmp/agents.db"),
8    learning=True,
9)

One line. Your agent now remembers users and improves over time.

Learning Stores

Each store captures a different type of knowledge:

Store	What it captures	Scope
User Profile	Structured facts (name, role, preferences)	Per user
User Memory	Unstructured observations from conversations	Per user
Session Context	Goals, plans, and progress for the current session	Per session
Entity Memory	Facts about external things (companies, projects, people)	Configurable
Learned Knowledge	Insights that transfer across users	Configurable
Decision Log	Decisions with reasoning for auditing and learning	Per agent

Stores can be enabled individually or combined. Each implements the same protocol: recall, process, build_context, get_tools.

Learning Modes

Control how and when the agent learns:

Mode	How it works
Always	Extraction runs automatically after each response
Agentic	Agent receives tools and decides what to save
Propose	Agent proposes learnings, you approve before saving

Namespaces

Some stores support configurable sharing scope via namespace:

Namespace	Who can access
`"user"`	Only the current user
`"global"`	Everyone (default)
Custom	Explicit grouping (e.g., `"engineering"`, `"sales_west"`)

Maintenance

The Curator keeps memories healthy:

1lm = agent.get_learning_machine()
2
3# Remove memories older than 90 days
4lm.curator.prune(user_id="alice", max_age_days=90)
5
6# Remove duplicates
7lm.curator.deduplicate(user_id="alice")