Thanks for asking question! Assuming you are using azure media services.
Video Indexer lets you create custom Language models to customize speech recognition by uploading adaptation text, namely text from the domain whose vocabulary you'd like the engine to adapt to. Once you train your model, new words appearing in the adaptation text will be recognized.
For a detailed overview and best practices for custom language models, see Customize a Language model with Video Indexer
You can use the Video Indexer website to create and edit custom Language models in your account.
Check official doc : Customize a Language model with the Video Indexer API