How to Intermediate · 4 min read

How to send video to Gemini API

Q: How to send video to Gemini API

To send a video to the Gemini API, encode the video file as base64 or provide a URL, then include it in the messages or inputs parameter depending on the API endpoint. Use the gemini-1.5-flash or gemini-2.0-flash model with the official Google AI Python client, passing the video data as a binary or base64 string in the request payload.

Quick answer

To send a video to the Gemini API, encode the video file as base64 or provide a URL, then include it in the messages or inputs parameter depending on the API endpoint. Use the gemini-1.5-flash or gemini-2.0-flash model with the official Google AI Python client, passing the video data as a binary or base64 string in the request payload.

PREREQUISITES

Python 3.8+
Google Cloud account with Gemini API access
pip install google-ai==0.1.0 or latest Google AI SDK
Set environment variable GOOGLE_API_KEY with your API key

Setup

Install the official Google AI Python SDK and set your API key as an environment variable.

Run pip install google-ai to install the SDK.
Export your API key: export GOOGLE_API_KEY='your_api_key_here' on Linux/macOS or set it in your environment variables on Windows.

bash

pip install google-ai

Step by step

Use the Google AI SDK to send a video file to the Gemini API by reading the video as bytes and encoding it in base64. Then include it in the messages payload with the appropriate content type.

python

import os
import base64
from google.ai import AIClient

# Initialize client with API key from environment
client = AIClient(api_key=os.environ["GOOGLE_API_KEY"])

# Load video file and encode as base64
video_path = "path/to/video.mp4"
with open(video_path, "rb") as f:
    video_bytes = f.read()
video_b64 = base64.b64encode(video_bytes).decode("utf-8")

# Prepare message with video content
messages = [
    {
        "role": "user",
        "content": {
            "type": "video_base64",
            "data": video_b64,
            "mime_type": "video/mp4"
        }
    }
]

# Call Gemini model
response = client.chat.completions.create(
    model="gemini-1.5-flash",
    messages=messages
)

print(response.choices[0].message.content)

output

Video processed successfully. Response text here...

Common variations

You can also send videos by URL instead of base64 encoding. Use the type field as video_url and provide the url key in the content dictionary.

For asynchronous or streaming use cases, the Google AI SDK supports async calls and streaming completions with similar parameters.

python

messages_url = [
    {
        "role": "user",
        "content": {
            "type": "video_url",
            "url": "https://example.com/video.mp4",
            "mime_type": "video/mp4"
        }
    }
]

response_url = client.chat.completions.create(
    model="gemini-2.0-flash",
    messages=messages_url
)

print(response_url.choices[0].message.content)

output

Video URL processed successfully. Response text here...

Troubleshooting

If you get a 400 Bad Request, verify your video encoding and MIME type.
For 401 Unauthorized, check your API key environment variable.
Large videos may require chunking or using URLs instead of base64.

✅

Key Takeaways

Use base64 encoding or video URLs to send video content to Gemini API.
Set your API key securely in environment variables before calling the API.
Choose the appropriate Gemini model like gemini-1.5-flash or gemini-2.0-flash for video input.
Check MIME types and encoding format to avoid request errors.
For large videos, prefer URLs or chunked uploads over base64 encoding.

Verified 2026-04 · gemini-1.5-flash, gemini-2.0-flash

Verify ↗