Indexing is the step that turns a file into something you can actually find. Until an asset has been indexed, it is searchable by its name and its metadata. After indexing, it is searchable by what is inside it.

ioMoVo runs the asset through the AI models you have connected and attaches what they produce back to the asset. Depending on the content and the models, that can mean text pulled out of a document or a scan, a transcript of spoken audio, topics and subjects detected in an image or a video, and classification against your taxonomy.

The practical difference is large. An unindexed archive answers "find the file called Q3_final_v2". An indexed one answers "find the shot where someone is talking about the harbour".

You can send one asset for indexing, or select a folder or a set of assets and send them together. For a new library, bulk is what you want. For a single file that just arrived, the individual route is quicker.

Which models you have connected. Indexing can only do what your connected models support. If transcription is not turning up, the usual cause is that no speech to text model is attached, not that indexing failed.

The model type. Every model carries a type, and that type gates which capabilities can run against it. See Connecting your own AI models if something you expect is unavailable.

Indexing is real processing work. A large first pass across an existing library takes time and, because you are using your own model keys, it bills to your own provider account.

Run a small representative sample first. You will learn what the results look like and what a full pass will cost, and both are much easier to act on before you commit than after.

If you connect a better model or add a capability later, you can send assets through again to pick up the new enrichment. Existing content does not stay stuck with whatever was available the first time.

Connecting your own AI models covers attaching models. What ioReports shows you covers confirming indexing completed across a set.