Skip to main content
The Session Context Store captures the current state of a conversation - what’s been discussed, what the goal is, and what progress has been made. Unlike other stores that accumulate data, session context is a snapshot that gets replaced on each update.

​
Overview

AspectDetails
Data TypeSummary, goal, plan, progress
ScopePer session (identified by session_id)
PersistenceSession lifetime (replaced on update)
Default ModeALWAYS
Supported ModesALWAYS only

​
Basic Usage

from agno.agent import Agent
from agno.db.postgres import PostgresDb
from agno.learn import LearningMachine
from agno.models.openai import OpenAIResponses

agent = Agent(
    model=OpenAIResponses(id="gpt-5.2"),
    db=PostgresDb(db_url="postgresql+psycopg://ai:ai@localhost:5532/ai"),
    learning=LearningMachine(
        session_context=True,
    ),
)

# Session tracks what's being discussed
agent.print_response(
    "I'm designing a REST API for a todo app. Should I use PUT or PATCH for updates?",
    user_id="[email protected]",
    session_id="api_design",
)

# Later in the session, context is maintained
agent.print_response(
    "What about the delete endpoint?",
    user_id="[email protected]",
    session_id="api_design",
)
# Agent knows the ongoing context about REST API design

​
Summary Mode (Default)

Summary mode captures the essence of the conversation without detailed planning: 
from agno.learn import LearningMachine, SessionContextConfig

agent = Agent(
    model=OpenAIResponses(id="gpt-5.2"),
    db=db,
    learning=LearningMachine(
        session_context=SessionContextConfig(),  # Summary mode by default
    ),
)
What gets captured:
  • What’s being worked on and why
  • Key decisions made and their rationale
  • Current state of any work in progress
  • Open questions or pending items

​
Planning Mode

Enable planning mode to track goals, plans, and progress: 
from agno.learn import LearningMachine, SessionContextConfig

agent = Agent(
    model=OpenAIResponses(id="gpt-5.2"),
    db=db,
    learning=LearningMachine(
        session_context=SessionContextConfig(
            enable_planning=True,
        ),
    ),
)

# Agent tracks goal, plan steps, and progress
agent.print_response(
    "Help me deploy a Python app to production. Give me the steps.",
    user_id="[email protected]",
    session_id="deploy_app",
)

# Later, progress is tracked
agent.print_response(
    "Done with step 1. What's next?",
    user_id="[email protected]",
    session_id="deploy_app",
)

​
Session Context Data Model

Each session context contains:
FieldTypeDescription
session_idstrUnique session identifier
user_idstrUser this session belongs to
summarystrWhat’s been discussed in this session
goalstrWhat user is trying to accomplish (planning mode)
planlistSteps to achieve goal (planning mode)
progresslistCompleted steps (planning mode)
created_atdatetimeWhen the context was created
updated_atdatetimeLast update timestamp

​
Accessing Session Context

# Get the learning machine
learning = agent.get_learning_machine()

# Access the session context store
context = learning.session_context_store.get(session_id="api_design")

if context:
    print(f"Summary: {context.summary}")
    if context.goal:
        print(f"Goal: {context.goal}")
    if context.plan:
        print(f"Plan: {context.plan}")
    if context.progress:
        print(f"Progress: {context.progress}")

# Debug output
learning.session_context_store.print(session_id="api_design")

​
Context Injection

Session context is automatically injected into the agent’s system prompt: 
<session_context>
This is a continuation of an ongoing session. Here's where things stand:

**Summary:** Helping user design a REST API for a todo app. Discussed resource naming conventions and decided on /todos endpoint. Currently exploring HTTP methods for CRUD operations.

**Current Goal:** Design complete REST API for todo application

**Plan:**
  1. Define resource endpoints
  2. Choose HTTP methods for each operation
  3. Design request/response schemas
  4. Add authentication

**Completed:**
  ✓ Define resource endpoints
</session_context>

​
Why Session Context Matters

Session context is essential when:
  1. Message history gets truncated - Long conversations may lose early context
  2. Sessions are resumed - User returns after a break
  3. Complex multi-step tasks - Track progress through long workflows
  4. Handoffs - Another agent or human needs to understand the state

​
Example: Technical Support Session

from agno.agent import Agent
from agno.db.postgres import PostgresDb
from agno.learn import LearningMachine, SessionContextConfig
from agno.models.openai import OpenAIResponses

agent = Agent(
    model=OpenAIResponses(id="gpt-5.2"),
    db=PostgresDb(db_url="postgresql+psycopg://ai:ai@localhost:5532/ai"),
    learning=LearningMachine(
        session_context=SessionContextConfig(enable_planning=True),
    ),
    instructions="""You are a technical support agent. Track the user's issue
    and your troubleshooting progress. When you identify the problem,
    create a plan to resolve it.""",
)

# Session 1: Initial issue report
agent.print_response(
    "My application keeps crashing when I try to upload files larger than 10MB.",
    user_id="[email protected]",
    session_id="support_ticket_123",
)

# Session 2: Troubleshooting continues
agent.print_response(
    "I checked the logs - there's an OutOfMemoryError.",
    user_id="[email protected]",
    session_id="support_ticket_123",
)
# Agent has full context of the ongoing troubleshooting session

​
Combining with Other Stores

Session context works well alongside user-level stores: 
from agno.learn import LearningMachine, UserProfileConfig, SessionContextConfig

agent = Agent(
    model=OpenAIResponses(id="gpt-5.2"),
    db=db,
    learning=LearningMachine(
        user_profile=UserProfileConfig(),     # Who the user is
        session_context=SessionContextConfig( # Current session state
            enable_planning=True,
        ),
    ),
)
This gives the agent both long-term user knowledge and short-term session state.