Postgres Tools

Inspect and query a Postgres database with PostgresTools, including read-only tool filtering.

postgres_tools.py
"""
Postgres Tools
=============================

Demonstrates postgres tools.
"""

from agno.agent import Agent
from agno.tools.postgres import PostgresTools

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


# Example 1: Include specific Postgres functions (default behavior - all functions included)
agent = Agent(
    tools=[
        PostgresTools(
            host="localhost",
            port=5532,
            db_name="ai",
            user="ai",
            password="ai",
            table_schema="ai",
        )
    ]
)

# Example 2: Include only read-only operations
agent_readonly = Agent(
    tools=[
        PostgresTools(
            host="localhost",
            port=5532,
            db_name="ai",
            user="ai",
            password="ai",
            table_schema="ai",
            include_tools=[
                "show_tables",
                "describe_table",
                "summarize_table",
                "inspect_query",
            ],
        )
    ]
)

# Example 3: Exclude dangerous operations
agent_safe = Agent(
    tools=[
        PostgresTools(
            host="localhost",
            port=5532,
            db_name="ai",
            user="ai",
            password="ai",
            table_schema="ai",
            exclude_tools=["run_query"],  # Exclude direct query execution
        )
    ]
)

# ---------------------------------------------------------------------------
# Run Agent
# ---------------------------------------------------------------------------
if __name__ == "__main__":
    agent.print_response(
        "List the tables in the database and summarize one of the tables", markdown=True
    )

    agent.print_response("""
    Please run a SQL query to get all sessions in `agno_sessions` created in the last 24 hours and summarize the table.
    """)

Connect to an existing database

Update the connection settings in each toolkit to a reachable Postgres database. The example expects an existing ai schema and a populated agno_sessions table; installing the packages does not create either. For a different database, change table_schema and the session-query prompt to match its tables. Ask the agent to inspect the columns before constructing a time filter.

Only agent runs in this script. To try a filtered configuration, change the calls to agent_readonly or agent_safe. New connections created by PostgresTools use read-only transactions; an explicitly supplied connection retains its existing settings. Excluding run_query still leaves export_table_to_path, which writes a local CSV. The explicit include_tools list also excludes that export function.

Run the Example

Set up your virtual environment

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

Install dependencies

uv pip install -U agno openai psycopg psycopg-binary

Export your OpenAI API key

export OPENAI_API_KEY="your_openai_api_key_here"

Run the example

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

python postgres_tools.py

Full source: cookbook/91_tools/postgres_tools.py