Meeting Summarizer Agent

"""Example: Meeting Summarizer & Visualizer Agent

This script uses OpenAITools (transcribe_audio, generate_image, generate_speech)
to process a meeting recording, summarize it, visualize it, and create an audio summary.

Requires: pip install openai agno
"""

import base64
from pathlib import Path
from textwrap import dedent

from agno.agent import Agent
from agno.models.google import Gemini
from agno.tools.openai import OpenAITools
from agno.tools.reasoning import ReasoningTools
from agno.utils.media import download_file, save_base64_data

input_audio_url: str = (
    "https://agno-public.s3.us-east-1.amazonaws.com/demo_data/sample_audio.mp3"
)

local_audio_path = Path("tmp/meeting_recording.mp3")
print(f"Downloading file to local path: {local_audio_path}")
download_file(input_audio_url, local_audio_path)

meeting_agent: Agent = Agent(
    model=Gemini(id="gemini-2.0-flash"),
    tools=[OpenAITools(), ReasoningTools()],
    description=dedent("""\
        You are an efficient Meeting Assistant AI.
        Your purpose is to process audio recordings of meetings, extract key information,
        create a visual representation, and provide an audio summary.
    """),
    instructions=dedent("""\
        Follow these steps precisely:
        1. Receive the path to an audio file.
        2. Use the `transcribe_audio` tool to get the text transcription.
        3. Analyze the transcription and write a concise summary highlighting key discussion points, decisions, and action items.
        4. Based *only* on the summary created in step 3, generating important meeting points. This should be a essentially an overview of the summary's content properly ordered and formatted in the form of meeting minutes.
        5. Convert the meeting minutes into an audio summary using the `generate_speech` tool.
    """),
    markdown=True,
)

response = meeting_agent.run(
    f"Please process the meeting recording located at '{local_audio_path}'"
)
if response.audio:
    base64_audio = base64.b64encode(response.audio[0].content).decode("utf-8")
    save_base64_data(base64_audio, Path("tmp/meeting_summary.mp3"))
    print(f"Meeting summary saved to: {Path('tmp/meeting_summary.mp3')}")