Postgres for Agent

PostgresDb stores an Agent's sessions and run history in PostgreSQL.

Usage

Install dependencies:

uv pip install agno openai ddgs sqlalchemy "psycopg[binary]"

Export your OpenAI API key:

export OPENAI_API_KEY="your_openai_api_key_here"

Run PostgreSQL

Install Docker Desktop, then start PostgreSQL with pgvector on port 5532:

docker run -d \
  -e POSTGRES_DB=ai \
  -e POSTGRES_USER=ai \
  -e POSTGRES_PASSWORD=ai \
  -e PGDATA=/var/lib/postgresql \
  -v pgvolume:/var/lib/postgresql \
  -p 5532:5432 \
  --name pgvector \
  agnohq/pgvector:18

postgres_for_agent.py

from agno.agent import Agent
from agno.db.postgres import PostgresDb
from agno.tools.websearch import WebSearchTools

db_url = "postgresql+psycopg://ai:ai@localhost:5532/ai"

db = PostgresDb(db_url=db_url)

agent = Agent(
    db=db,
    tools=[WebSearchTools()],
    add_history_to_context=True,
)
agent.print_response("How many people live in Canada?")
agent.print_response("What is their national anthem called?")

Run the example

Save the code as postgres_for_agent.py, start the database described above, and run:

python postgres_for_agent.py

Parameters

Parameter	Type	Default	Description
`id`	`Optional[str]`	-	The ID of the database instance. UUID by default.
`db_url`	`Optional[str]`	-	The database URL to connect to.
`db_engine`	`Optional[Engine]`	-	The SQLAlchemy database engine to use.
`db_schema`	`Optional[str]`	-	The database schema to use.
`session_table`	`Optional[str]`	-	Name of the table to store Agent, Team and Workflow sessions.
`runs_table`	`Optional[str]`	-	Name of the table to store the runs of each session.
`memory_table`	`Optional[str]`	-	Name of the table to store memories.
`metrics_table`	`Optional[str]`	-	Name of the table to store metrics.
`eval_table`	`Optional[str]`	-	Name of the table to store evaluation runs data.
`knowledge_table`	`Optional[str]`	-	Name of the table to store knowledge content.
`traces_table`	`Optional[str]`	-	Name of the table to store traces.
`spans_table`	`Optional[str]`	-	Name of the table to store spans.
`versions_table`	`Optional[str]`	-	Name of the table to store schema versions.
`components_table`	`Optional[str]`	-	Name of the table to store components.
`component_configs_table`	`Optional[str]`	-	Name of the table to store component configurations.
`component_links_table`	`Optional[str]`	-	Name of the table to store links between components.
`learnings_table`	`Optional[str]`	-	Name of the table to store learnings.
`schedules_table`	`Optional[str]`	-	Name of the table to store cron schedules.
`schedule_runs_table`	`Optional[str]`	-	Name of the table to store schedule run history.
`job_table`	`Optional[str]`	-	Name of the table to store durable background run jobs.
`approvals_table`	`Optional[str]`	-	Name of the table to store human approval requests.
`auth_tokens_table`	`Optional[str]`	-	Name of the table to store OAuth tokens for external services.
`service_accounts_table`	`Optional[str]`	-	Name of the table to store service accounts.
`mcp_oauth_clients_table`	`Optional[str]`	-	Name of the table to store MCP OAuth client registrations.
`mcp_oauth_transactions_table`	`Optional[str]`	-	Name of the table to store MCP OAuth transactions.
`mcp_oauth_codes_table`	`Optional[str]`	-	Name of the table to store MCP OAuth authorization codes.
`mcp_oauth_refresh_tokens_table`	`Optional[str]`	-	Name of the table to store MCP OAuth refresh tokens.
`mcp_oauth_keys_table`	`Optional[str]`	-	Name of the table to store MCP OAuth signing keys.
`create_schema`	`bool`	`True`	Whether to create the database schema if it doesn't exist. Set to `False` when the schema is managed externally.

Usage

Run PostgreSQL

Run the example

Parameters

On this page