Claude Model Test

Tests learning with Claude instead of OpenAI.

claude_model.py

"""
Claude Model Test
=================
Tests learning with Claude instead of OpenAI.

All other cookbooks use OpenAI (gpt-5.5). This test verifies that
learning works with Claude models, ensuring the implementation is
model-agnostic.

Key things to verify:
1. Profile extraction works with Claude
2. Tool calls work correctly (Claude uses different tool format)
3. Background extraction completes successfully
"""

from agno.agent import Agent
from agno.db.postgres import PostgresDb
from agno.learn import LearningMachine, LearningMode, UserProfileConfig
from agno.models.anthropic import Claude

# ---------------------------------------------------------------------------
# Create Agent - Using Claude instead of OpenAI
# ---------------------------------------------------------------------------

db = PostgresDb(db_url="postgresql+psycopg://ai:ai@localhost:5532/ai")

agent = Agent(
    model=Claude(id="claude-sonnet-4-6"),  # Using Claude
    db=db,
    learning=LearningMachine(
        user_profile=UserProfileConfig(
            mode=LearningMode.ALWAYS,
        ),
    ),
    markdown=True,
)

# ---------------------------------------------------------------------------
# Run Demo
# ---------------------------------------------------------------------------

if __name__ == "__main__":
    user_id = "[e-mail]"

    print("\n" + "=" * 60)
    print("TEST: Learning with Claude model")
    print("=" * 60 + "\n")

    print(f"Model type: {type(agent.model).__name__}")

    # Session 1: Share information
    print("\n" + "=" * 60)
    print("SESSION 1: Share information (Claude extraction)")
    print("=" * 60 + "\n")

    agent.print_response(
        "Hi! I'm Bruce Wayne, but my friends call me Batman.",
        user_id=user_id,
        session_id="claude_session_1",
        stream=True,
    )

    # Check if LearningMachine was initialized
    lm = agent.learning_machine
    print(f"\nLearningMachine exists: {lm is not None}")

    if lm and lm.user_profile_store:
        lm.user_profile_store.print(user_id=user_id)
    else:
        print("\n[WARNING] UserProfileStore not available - extraction may have failed")
        print(
            "Note: Some Claude models may not support structured outputs required for extraction"
        )

    # Session 2: Verify profile persisted
    print("\n" + "=" * 60)
    print("SESSION 2: Profile recall (Claude)")
    print("=" * 60 + "\n")

    agent.print_response(
        "What's my secret identity?",
        user_id=user_id,
        session_id="claude_session_2",
        stream=True,
    )

    if lm and lm.user_profile_store:
        lm.user_profile_store.print(user_id=user_id)

    print("\n" + "=" * 60)
    print("CLAUDE MODEL TEST COMPLETE")
    print("=" * 60)

Run the Example

1. Set up your virtual environment

   Mac

   uv venv --python 3.12
   source .venv/bin/activate
   

   uv venv --python 3.12
   .venv\Scripts\activate
   
2. Install dependencies

   uv pip install -U agno anthropic psycopg-binary sqlalchemy
   
3. Export your Anthropic API key

   Mac/Linux

   export ANTHROPIC_API_KEY="your_anthropic_api_key_here"
   

   $Env:ANTHROPIC_API_KEY="your_anthropic_api_key_here"
   
4. Run PgVector

   docker run -d \
     -e POSTGRES_DB=ai \
     -e POSTGRES_USER=ai \
     -e POSTGRES_PASSWORD=ai \
     -e PGDATA=/var/lib/postgresql/data/pgdata \
     -v pgvolume:/var/lib/postgresql/data \
     -p 5532:5432 \
     --name pgvector \
     agnohq/pgvector:18
   
5. Run the example

   Save the code above as claude_model.py, then run:

   python claude_model.py
   

Full source: cookbook/08_learning/06_quick_tests/04_claude_model.py