Gemini Interactions - Multi-turn Conversation

Demonstrates server-side conversation history with the Interactions API.

After the first response, subsequent turns send the new message and reference the previous interaction via previous_interaction_id. This example demonstrates conversation continuity without resending the full request history. Cost savings depend on an implicit-cache hit and the model’s eligibility requirements; these short prompts do not demonstrate savings. See Google caching.

multi_turn.py
"""
Gemini Interactions - Multi-turn Conversation
==============================================

Demonstrates server-side conversation history with the Interactions API.
After the first response, subsequent turns only send the new message
and reference the previous interaction via `previous_interaction_id`.
This enables implicit caching and reduces token costs.

Multi-turn requires a db (e.g. SqliteDb) so the interaction_id from each
turn's response is persisted on the assistant message and read back on
the next turn.
"""

from agno.agent import Agent
from agno.db.sqlite import SqliteDb
from agno.models.google import GeminiInteractions

# ---------------------------------------------------------------------------
# Create Agent
# ---------------------------------------------------------------------------

agent = Agent(
    model=GeminiInteractions(id="gemini-3.7-flash"),
    add_history_to_context=True,
    db=SqliteDb(db_file="tmp/data.db"),
    markdown=True,
)

# ---------------------------------------------------------------------------
# Run Agent
# ---------------------------------------------------------------------------
if __name__ == "__main__":
    # First turn - establishes the interaction
    agent.print_response("My name is Alice and I love hiking in the mountains.")

    # Second turn - references the previous interaction for context
    agent.print_response("What did I just tell you about myself?")

    # Third turn - continues the conversation chain
    agent.print_response(
        "Suggest a hiking destination based on what you know about me."
    )

Run the Example

Set up your virtual environment

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

Install dependencies

uv pip install -U agno google-genai sqlalchemy

Export your Google API key

export GOOGLE_API_KEY="your_google_api_key_here"

Run the example

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

python multi_turn.py

Full source: cookbook/90_models/google/gemini_interactions/multi_turn.py