Asset Processors
Overview of the Nomad Media asset processing pipeline — background Lambda functions that analyze, enrich, and index assets automatically on upload.
When an asset is uploaded to a Nomad Media S3 bucket, a background pipeline of Lambda functions automatically processes it. These processors run in parallel where possible to minimize total processing time, and their output enriches the asset manifest with structured metadata used by search, playback, and content intelligence features.
In This Section
| Page | Description |
|---|---|
| Asset Processor Overview | Pipeline architecture, how processors are triggered, and parallel execution model |
| Auto-Attach Related Assets | Automatically attach thumbnails, previews, subtitles, and metadata files based on file naming conventions |
| Format Support | Supported image, video, and audio input formats |
| Metadata Folders and Structures | How and where processed metadata is stored (metadata bucket vs. metadata-archive bucket) |
| Reprocessing Assets | How to trigger reprocessing of a single asset or folder, including selective processor overrides |
| Turning On/Off Asset Processors and Dependencies | Enabling/disabling processors via config, URL match/exclude rules, and processor dependency chains |
| Custom Asset Processors | Building a custom Lambda processor: Lambda types, data structures, Redis state, and deployment |
| Textract vs. RekognitionImageDetectText | Comparison of Amazon Textract and Rekognition for text detection across different image types |
| Rekognition IndexFaces and FaceIDs | How Rekognition IndexFaces assigns FaceIDs with and without ExternalImageID |
| Rekognition Face Detection and Matching | DynamoDB table schema and workflow for face detection, matching, and person creation |
How the Pipeline Works
Assets dropped into a configured S3 bucket trigger the processing pipeline automatically. The pipeline:
- Detects the asset type (video, image, audio, document)
- Runs applicable processors in parallel based on the
application/processorsconfig - Writes output to the metadata bucket (public-facing) or metadata-archive bucket (internal)
- Assembles a final asset manifest once all processors complete
- Registers the manifest and related assets via the Nomad Media Admin API
Processors can be individually enabled, disabled, or scoped to specific S3 path prefixes using URL match/exclude rules. See Turning On/Off Asset Processors for details.