Rules Engine Overview

Every processing decision in Nomad Media — proxy generation, transcoding, AI enrichment, archiving — is governed by a single unified rules engine. Once rules are configured, the system runs fully unattended: no manual steps are required when new content arrives.

How It Works

When a file lands in your S3 bucket, Nomad Media evaluates the configured rules for each processor and decides whether that processor should run. Rules can target content in two ways:

• By folder — using URL path prefix matching (urlMatches / urlExcludes)
• By asset properties — using expression-based conditions evaluated against the file's metadata

These two targeting methods can be combined within a single rule for precise control.

Folder Targeting

The simplest form of a rule is a folder match. A rule with urlMatches: ["content/radio/"] applies only to files in that folder and its subfolders. Specific subfolders can be excluded using urlExcludes.

This makes it straightforward to apply different processing behavior to different parts of your library. Examples:

• Apply AI transcription only to the content/published/ folder
• Apply Glacier archiving rules to content/raw/ but exclude content/raw/active/
• Generate HLS streaming renditions for content/video/broadcast/ but not content/video/rough-cuts/

Expression-Based Conditions

Rules can also evaluate properties of the asset itself — such as resolution, bitrate, duration, file extension, or media type — using a simple expression syntax.

videoWidth > 1920 or videoHeight > 1080
videoBitRate > 6291456
fileExtension == '.mp4'
audioChannels > 2
contentLength > 5242880

This allows the same folder to contain multiple asset types while only triggering specific processing for the right files. For example, a single ingest folder can contain both video and audio files, with video proxying triggered on video files and audio transcription triggered on audio files independently.

Rules Are Additive

Multiple rules can exist for the same processor, each targeting different folders or conditions. When a new file arrives, the system evaluates all active rules and runs every processor whose conditions are satisfied.

Rules can be added or updated at any time — including retroactively. When you add a new rule (for example, to enable AI processing on a folder that previously had none), you can trigger a reprocess on that folder and the system will add the missing outputs without regenerating anything that already exists.

Everything Uses the Same Engine

The same rule structure governs all of Nomad's processing capabilities:

Nothing in the system is a global on/off switch for the entire library. Every capability can be scoped to specific folders and file types.

Related Pages

• Turning On/Off Asset Processors and Dependencies — full JSON configuration reference for processor rules
• Proxy Generation Overview — how proxy rules are configured
• AI Enablement Guide — scoping AI processors by folder and file type
• Phased AI Rollout — how to stage AI by folder, tier, and test environment
• Media Archiving & Storage Lifecycle — archiving rules configuration
• Rule Syntax Reference — complete expression syntax documentation