Skip to main content

​
Code

workflow.py
from agno.agent import Agent
from agno.db.sqlite import SqliteDb
from agno.models.openai import OpenAIChat
from agno.os.app import AgentOS
from agno.os.interfaces.telegram import Telegram
from agno.workflow.step import Step
from agno.workflow.steps import Steps
from agno.workflow.workflow import Workflow

agent_db = SqliteDb(
    session_table="telegram_workflow_sessions", db_file="tmp/telegram_workflow.db"
)

drafter = Agent(
    name="Drafter",
    model=OpenAIChat(id="gpt-4o-mini"),
    instructions="Draft a response to the user's message. Be helpful and informative.",
)

editor = Agent(
    name="Editor",
    model=OpenAIChat(id="gpt-4o-mini"),
    instructions=[
        "Review and polish the draft for clarity and conciseness.",
        "Keep it short and suitable for a Telegram message.",
    ],
)

draft_step = Step(
    name="draft",
    agent=drafter,
    description="Draft an initial response",
)

edit_step = Step(
    name="edit",
    agent=editor,
    description="Edit and polish the draft",
)

telegram_workflow = Workflow(
    name="Telegram Draft-Edit Workflow",
    description="A two-step workflow that drafts and then edits responses for Telegram",
    steps=[
        Steps(
            name="draft_and_edit",
            description="Draft then edit a response",
            steps=[draft_step, edit_step],
        )
    ],
    db=agent_db,
)

agent_os = AgentOS(
    workflows=[telegram_workflow],
    interfaces=[
        Telegram(
            workflow=telegram_workflow,
            reply_to_mentions_only=True,
        )
    ],
)
app = agent_os.get_app()

if __name__ == "__main__":
    agent_os.serve(app="workflow:app", reload=True)

​
Usage

1

Set up your virtual environment

uv venv --python 3.12
source .venv/bin/activate
2

Set Environment Variables

export TELEGRAM_TOKEN=your-bot-token-from-botfather
export OPENAI_API_KEY=your-openai-api-key
export APP_ENV=development
3

Install dependencies

uv pip install -U "agno[telegram]"
4

Run Example

python workflow.py

​
Key Features

  • Two-Step Pipeline: Drafter writes an initial response, Editor polishes it
  • Workflow on Telegram: The Workflow (not individual agents) is passed to the Telegram interface
  • Sequential Steps: Steps chains Step objects in order
  • Persistent Sessions: SQLite database for conversation history across restarts