S3 URL File Input

Generate an S3 pre-signed URL with boto3 and pass it straight to Gemini to summarize the PDF without downloading it.

The Gemini API now supports external HTTPS URLs (up to 100MB). Generate a pre-signed URL from S3 and pass it directly to Gemini.

s3_url_file_input.py

"""
Example: Analyze files from AWS S3 using pre-signed URLs.

The Gemini API now supports external HTTPS URLs (up to 100MB).
Generate a pre-signed URL from S3 and pass it directly to Gemini.

Requirements:
- AWS credentials configured (via environment variables or ~/.aws/credentials)
- boto3 installed: uv pip install boto3

Supported formats: PDF, JSON, HTML, CSS, XML, images (PNG, JPEG, WebP, GIF)

Note: External URL support requires Gemini 3.x models (e.g., gemini-3.5-flash).
      Gemini 2.0 models do not support this feature.
"""

import boto3
from agno.agent import Agent
from agno.media import File
from agno.models.google import Gemini

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

# Generate a pre-signed URL for your S3 object
# Replace with your own bucket and key for private files
s3_client = boto3.client("s3")
presigned_url = s3_client.generate_presigned_url(
    "get_object",
    Params={
        "Bucket": "agno-public",  # Example: using Agno's public bucket
        "Key": "recipes/ThaiRecipes.pdf",
    },
    ExpiresIn=3600,  # URL valid for 1 hour
)

agent = Agent(
    model=Gemini(id="gemini-3.5-flash"),
    markdown=True,
)

# Pass pre-signed URL directly - Gemini fetches the content
agent.print_response(
    "What is this document about? Answer in one sentence.",
    files=[
        File(
            url=presigned_url,
            mime_type="application/pdf",
        )
    ],
)

# ---------------------------------------------------------------------------
# Run Agent
# ---------------------------------------------------------------------------

if __name__ == "__main__":
    pass

Run the Example

1. Set up your virtual environment

   Mac

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

   uv venv --python 3.12
   .venv\Scripts\activate
   
2. Install dependencies

   uv pip install -U agno boto3 google-genai
   
3. Export your Google API key

   Mac/Linux

   export GOOGLE_API_KEY="your_google_api_key_here"
   

   $Env:GOOGLE_API_KEY="your_google_api_key_here"
   
4. Configure AWS credentials

   Configure boto3 through environment variables, ~/.aws/credentials, or an IAM role. The source presigns s3://agno-public/recipes/ThaiRecipes.pdf; if you use another object, update the bucket and key and give the AWS identity permission to read it.

5. Run the example

   Save the code above as s3_url_file_input.py, then run:

   python s3_url_file_input.py
   

Full source: cookbook/90_models/google/gemini/s3_url_file_input.py