File uploads missing Content-Type headers across all storage providers (S3, GCP, Supabase)

[hamchowderr]

Description

The NCAT API uploads files to cloud storage without specifying Content-Type headers across
all storage providers (S3-compatible and Google Cloud Storage). This causes upload failures
with Supabase Storage and prevents inline media viewing in browsers.

Environment

• NCAT Version: stephengpope/no-code-architects-toolkit:latest
• Build Number: 206
• Python Version: 3.9.23
• Deployment: Docker
• Tested Storage Backends:
  • Supabase Storage (S3-compatible) ❌ Fails with InvalidMimeType
  • AWS S3 ⚠️ Defaults to application/octet-stream
  • DigitalOcean Spaces ⚠️ Defaults to application/octet-stream
  • Google Cloud Storage ⚠️ Auto-detects (unreliable)

Problem Details

Error from production logs:
ERROR:services.s3_toolkit:Error uploading file to S3: An error occurred (InvalidMimeType)
when calling the PutObject operation: mime type application/octet-stream is not supported

Failure Statistics (Oct 15-21, 2025):

• 42 unique job failures over 7 days
• 41 failures today alone (accelerating issue)
• Affected endpoints:
  • /v1/image/convert/video - 41 failures (98%)
  • /v1/media/download - 1 failure (2%)

User Impact:

1. ❌ Supabase Storage rejects uploads entirely
2. ❌ MP4 videos download instead of playing in browser
3. ❌ MP3 audio files download instead of streaming
4. ❌ Images download instead of displaying inline

Root Cause

1. S3-Compatible Provider (services/s3_toolkit.py)

def upload_to_s3(file_path, s3_url, access_key, secret_key, bucket_name, region):
    # ...
    client.upload_fileobj(data, bucket_name,
        os.path.basename(file_path),
        ExtraArgs={'ACL': 'public-read'})  # ❌ No ContentType

Problem: boto3 defaults to application/octet-stream when ContentType is not specified.

2. Google Cloud Storage Provider (services/gcp_toolkit.py)

def upload_to_gcs(file_path, bucket_name=GCP_BUCKET_NAME):
    # ...
    blob = bucket.blob(os.path.basename(file_path))
    blob.upload_from_filename(file_path)  # ❌ No content_type parameter

Problem: GCP auto-detects MIME type, but this is unreliable.

Proposed Solution

Use Python's built-in mimetypes library to auto-detect Content-Type from file extensions:

Fix for services/s3_toolkit.py

import mimetypes

def upload_to_s3(file_path, s3_url, access_key, secret_key, bucket_name, region):
    session = boto3.Session(
        aws_access_key_id=access_key,
        aws_secret_access_key=secret_key,
        region_name=region
    )

    client = session.client('s3', endpoint_url=s3_url)

    try:
        # Auto-detect MIME type from file extension
        content_type, _ = mimetypes.guess_type(file_path)
        if not content_type:
            content_type = 'application/octet-stream'

        logger.info(f"Uploading {os.path.basename(file_path)} with Content-Type:
{content_type}")

        with open(file_path, 'rb') as data:
            client.upload_fileobj(data, bucket_name,
                os.path.basename(file_path),
                ExtraArgs={
                    'ACL': 'public-read',
                    'ContentType': content_type  # ✅ Add this
                })

        encoded_filename = quote(os.path.basename(file_path))
        file_url = f"{s3_url}/{bucket_name}/{encoded_filename}"
        return file_url
    except Exception as e:
        logger.error(f"Error uploading file to S3: {e}")
        raise

Fix for services/gcp_toolkit.py

import mimetypes

def upload_to_gcs(file_path, bucket_name=GCP_BUCKET_NAME):
    # ... existing validation ...

    # Auto-detect MIME type from file extension
    content_type, _ = mimetypes.guess_type(file_path)
    if not content_type:
        content_type = 'application/octet-stream'

    logger.info(f"Uploading {os.path.basename(file_path)} with Content-Type: {content_type}")

    bucket = gcs_client.bucket(bucket_name)
    blob = bucket.blob(os.path.basename(file_path))
    blob.upload_from_filename(file_path, content_type=content_type)  # ✅ Add content_type

    # ... rest of function ...

Benefits

✅ Eliminates InvalidMimeType errors with Supabase Storage✅ Enables inline viewing of media
files in browsers✅ Consistent behavior across all storage backends✅ Zero breaking changes -
fully backwards compatible✅ No new dependencies - uses Python standard library

Additional Notes

I'm happy to submit a PR with both fixes if this approach is acceptable. This issue affects all
 users regardless of which storage backend they choose.