Running With S3¶

Purpose¶

Use this flow when you want S3-backed inputs, request-selected S3 outputs, or both.

Build With The Right Feature¶

cargo run -p service-http --features s3

If S3 config exists but the binary is built without --features s3, startup fails.

Configure S3 Inputs¶

Enable [input_resolvers.s3] in config/service.toml:

[input_resolvers.s3]
region = "eu-central-1"
buckets = ["media-bucket", "publication-bucket"]

Configure S3 Outputs¶

Add [output_sinks.s3]:

[output_sinks.s3]
region = "eu-central-1"
buckets = ["processed-media", "archive-media"]
base_prefix = "exports"

The requested output.bucket must be in the configured allow-list. See the Configuration Guide for the full field reference.

Example Request¶

curl -X POST http://localhost:8080/v2/c2pa/video \
  -H "Content-Type: application/json" \
  -d '{
    "inputs": [
      {
        "type": "remote_file_url",
        "url": "s3://media-bucket/input/video.mp4",
        "filename_hint": "video.mp4"
      }
    ],
    "params": {},
    "output": {
      "type": "s3",
      "bucket": "processed-media",
      "prefix": "exports/c2pa",
      "name": "spring-campaign"
    }
  }'

The object key is base_prefix / output.prefix / output.name / <per-input-leaf>, e.g. exports/exports/c2pa/spring-campaign/video_c2pa.mp4. There is no job/batch id in the key, and reusing the same destination overwrites it.

Common Failures¶

startup error about missing feature: build with --features s3
validation failure for output bucket: add the bucket to output_sinks.s3.buckets
startup failure for bucket probe: verify credentials, region, endpoint, and bucket permissions