This is a simple Python/Flask demo of using the excellent jQuery File Upload to upload large files directly to S3 using browser uploads to S3 using HTML POST and S3 CORS. I largely followed the instruction in this Ruby example, but I thought other people may want to see the Python/Flask example.

To run the example on your computer, do something like the following:

git clone <repo url> /tmp/s3upload
cd /tmp/s3upload
virtualenv .venv
source .venv/bin/activate
pip install -r requirements.txt
export AWS_ACCESS_KEY_ID=<my amazon key ID>
export AWS_SECRET_ACCESS_KEY=<my amazon secret key>
export AWS_S3_BUCKET_URL=<bucket url>
python app.py

Naturally you should fill in the values in angle brackets with real values. AWS_S3_BUCKET_URL is the canonical URL of your bucket, for a bucket named foo in US East it would be https://foo.s3.amazonaws.com.

Your bucket will need to have a CORS policy set. Follow the instructions here to set a CORS policy. For real work, you'd like to have a policy confined to your origin domain and perhaps other conditions. For development, a suitable policy could be:

<CORSConfiguration>
    <CORSRule>
        <AllowedOrigin>*</AllowedOrigin>
        <AllowedMethod>GET</AllowedMethod>
        <AllowedMethod>POST</AllowedMethod>
        <AllowedHeader>*</AllowedHeader>
    </CORSRule>
</CORSConfiguration>

This is all you need to get the example working. I think the code is very straightforward, but for further explanation on what's going on I suggest you read the Ruby example linked above.

I'm @aknin by the way, contact me if you have any questions.