Inputs

An input connects a live resource to a stream. You define inputs as a list in spec.inputs, and each input specifies exactly one source of stream data. You choose exactly one of three fields per input. They are mutually exclusive, so you cannot mix them in a single input entry:

Field	Connects to	Use when
sourceName	One specific, named source resource (Media API)	You know exactly which physical feed you want, and it won’t change.
contentName	A content identifier, resolved to a candidate source that can supply it (Media API)	The physical source behind the feed might change, and you want the platform to resolve among available candidates.
flowName	Another live resource’s output	You are chaining resources together, feeding the output of one flow into another.

Before defining an input, retrieve the created config and copy its status.inputPins[].name. Replace <INPUT_PIN> in the following examples with a returned pin name.

A single camera with a direct source reference uses:

"inputs": [
  {
    "sourceName": "camera-1-source",
    "transformInput": "<INPUT_PIN>"
  }
]

A setup where several sources can supply the same logical content instead references their shared content name:

"inputs": [
  {
    "contentName": "camera-1",
    "transformInput": "<INPUT_PIN>"
  }
]

In both cases, the live resource keeps the same processing config. contentName lets the platform resolve among candidates, but it does not guarantee seamless media-aware failover.

Flow input refiners

When an input uses contentName, the platform has to pick one specific source (or another flow’s output) to actually supply it. See Resolving content for how that resolution works by default. A refiner narrows that choice further, without you having to switch to a hard-pinned sourceName or flowName:

"inputs": [
  {
    "contentName": "camera-1",
    "transformInput": "<INPUT_PIN>",
    "refiner": {
      "flow": { "name": "upstream-encode" }
    }
  }
]

There are two kinds of input refiner. flow restricts resolution to a single named upstream flow, the same “another live resource’s output” concept as the flowName field above, useful when several flows could technically supply the same content but you want this input fed by one specific one, without giving up the flexibility of contentName resolution entirely.

labelMatch restricts resolution to sources or flows carrying a specific label key and value, useful when you tag a group of resources and want an input to resolve to whichever one of them is currently reachable, rather than a single fixed name.

"refiner": {
  "labelMatch": { "key": "region", "value": "eu-west" }
}

Refiners only apply to inputs today. The Live API schema also defines a refiner field on outputs and on the flow as a whole, but both are currently empty placeholders reserved for future functionality, setting them has no effect yet.

What transform pins are

Every input has a transformInput field. This field names the slot, or pin, in the processing config that receives the stream. Single-input resources bind one input to one pin. Multiview resources bind several inputs to separate pins.

Think of pins as labelled input jacks on a mixing desk. Your encodingLive template defines how many inputs it expects and what each one is called. The transformInput value on each flow input tells the platform which jack to plug that camera feed into. If you send four camera feeds, each one gets routed to a different named pin, and the template decides how they are arranged on screen.

Get pin names

Pin names come from the config itself. After creating a config, retrieve it and read status.inputPins[].name. Use those exact values for transformInput. Presets and configs can change, so do not treat a pin name from an example as universal.

Check multiple bindings

Each pin also reports allowMultipleBindings. If it’s false, only one input can target that pin, a second input pointed at the same pin is a conflict, not a merge. If it’s true, the template has been built to accept more than one input on that pin. Check this before assuming you can route two feeds to what looks like the same slot.

Inspect your created config before assigning an input to a pin.

What comes next

• Outputs: where encoded output goes.
• Content, sources, and destinations: create the sources and content this page references.