IV has powerful in-built search and topic extraction capabilities, allowing you to search across words and word alternatives, and build powerful maps of how audio and video files relate to each other. Multi-tenanted, it lets you easily segment data, to handle different cases, departments etc. But many users already have access to search capabilities in their existing workflows, so using just the IV API, you can get all of the power of IV’s advanced text, alternatives, and topic capabilities, and put them direct into your current database.
IV has a “cloudy” API and architecture, but is powerful enough to be deployed on the edge for highly secure processing requirements. But for larger use cases, the system can easily be deployed into a data centre, or as part of an AWS or Azure strategy, allowing you to auto-scale resources as you need them.
Indexing voice for search can be a slow process if you are looking to extract grammatically correct phrases and sentence, rather than just indexing raw sounds. IV overcomes this with its pioneering AI strategy, which saw the first commercially available GPU speech processing in 2014. The high power and high density afforded by GPU servers allows you to index large volumes of audio files in the shortest time with the minimum amount of power and latency
Key to efficient retrieval is that your speech algorithms can find the words in the audio in the first place! IV’s “QuickTrain” methodology lets you the user train the system to find what is important to you using contextual data like search terms, names, emails or even Wikipedia. And you can now do your processing in 18 languages or dialects.