Daytona Speech Engine

Veridict Daytona speech engine features a speaker-independent technology that enables near 100% recognition accuracy out-of-the-box for most tasks, but which can be adapted to individual voices, dialects, etc. to improve recognition accuracy even further.

The engine is appropriate to applications from grammar-based command-and-control systems to demanding recognition and dictation tasks, covering natural language domains with large vocabularies. The software enforces no limits on vocabulary size or language model complexity; a standard desktop computer can provide real-time performance and can run n-gram language models with severeal millions of concurrent unique words.