Serve and embed

Put the data agent behind an API so a dashboard, a Slack channel, or a product widget can ask it questions.

A data agent that only runs in a notebook helps one analyst. Behind an API, it answers questions from a Slack channel, a BI dashboard’s natural-language box, or an “explain this metric” widget in your product. AgentOS turns the agent into that API.

Start PgVector on port 5532 before running this example.

data_agent.py

from agno.agent import Agent
from agno.db.postgres import PostgresDb
from agno.models.openai import OpenAIResponses
from agno.os import AgentOS
from agno.tools.sql import SQLTools

agent = Agent(
    id="data-agent",
    model=OpenAIResponses(id="gpt-5.5"),
    db=PostgresDb(db_url="postgresql+psycopg://ai:ai@localhost:5532/ai"),
    tools=[SQLTools(db_url="postgresql+psycopg://readonly@warehouse/analytics")],
    learning=True,
)

agent_os = AgentOS(agents=[agent])
app = agent_os.get_app()

if __name__ == "__main__":
    agent_os.serve(app="data_agent:app", port=7777)

Run it with python data_agent.py. The app="data_agent:app" import string has to match the module name. Every surface then calls the same endpoint:

curl -X POST http://localhost:7777/agents/data-agent/runs \
  -F 'message=MRR by plan, last 6 months' \
  -F 'user_id=[e-mail]' \
  -F 'session_id=q4-review' \
  -F 'stream=false'

Surfaces a data agent lives on

Surface	Shape
Slack channel	An interface maps a thread to a session; the team asks in plain English
Dashboard NL box	A widget posts the question and renders the answer plus its SQL
Scheduled digest	A cron job runs the agent and posts “yesterday’s numbers” every morning
Backend check	A pipeline calls the agent to sanity-check a metric before publishing

The serving model is identical to a product agent. The difference is what the agent does, not how it is served. user_id and session_id select the conversation history for a run.

This example leaves AgentOS authorization disabled. Callers can supply their own user and session identifiers. Enable AgentOS authorization before network exposure, and connect SQLTools with a database role that enforces read-only access.

Shared learnings, separate sessions

A data agent’s value compounds when corrections are shared across the team. Conversation threads stay scoped by user_id and session_id. learning=True only enables the per-user profile and memory stores, so to share what the agent learns, pass a knowledge base to LearningMachine instead. That enables the Learned Knowledge store, whose namespace defaults to "global", so a fix triggered by one analyst’s question helps everyone.

from agno.knowledge import Knowledge
from agno.learn import LearningMachine
from agno.vectordb.pgvector import PgVector

knowledge = Knowledge(
    vector_db=PgVector(
        db_url="postgresql+psycopg://ai:ai@localhost:5532/ai",
        table_name="warehouse_learnings",
    ),
)

agent = Agent(
    id="data-agent",
    # same model, db, and tools as above
    learning=LearningMachine(knowledge=knowledge),
)

The store runs in AGENTIC mode: the model decides when to call save_learning and search_learnings. To keep learnings inside one team or tenant, set namespace on LearningMachine.

State	Scope
Conversation thread	Per user_id and session_id
Learnings about the warehouse	Learned Knowledge store, shared "global" namespace by default

Next steps

Task	Guide
Add Slack or a browser surface	Interfaces
Lock down the endpoints	Security and auth

Developer Resources

• Serve as an API
• Product agents: serving