Conformer2

Overview

Improved automatic speech recognition with enhanced transcription.

Conformer-2: Conformer-2 is a cutting-edge AI model designed for automatic speech recognition. Trained on a massive 1.1 million hours of English audio data, Conformer-2 is the latest advancement building upon the success of Conformer-1. This model offers significant improvements in transcribing proper nouns, alphanumerics, and robustness to noise. By increasing the training data and implementing model ensembling techniques, Conformer-2 achieves remarkable improvements in error rates and processing speed compared to its predecessor. Its enhanced performance in real-world audio conditions makes it a valuable tool for developers creating speech-to-text applications. With an API available for testing and implementation, Conformer-2 allows users to leverage state-of-the-art speech recognition technology efficiently and effectively.

Use Cases

• Improve the accuracy of transcribing interviews and podcasts using Conformer-2's advanced speech recognition capabilities, ensuring accurate capture of proper nouns and alphanumerics for enhanced content understanding.
• Develop real-time voice-to-text applications for meetings and conferences with Conformer-2, benefitting from its robust noise handling abilities to maintain transcription quality in varying environmental conditions.
• Create automated subtitles for videos and films with precise transcription accuracy using Conformer-2, speeding up the subtitling process and increasing accessibility for viewers with hearing impairments.

Pricing

Free plan available. Paid plans unlock more usage.

• Speech-to-text: $0.37 second hour hour
• Real-time transcription: $0.47 second hour hour
• Audio intelligence: model price key phrases $0.01/hour sentiment analysis $0.02/hour summarization $0.03/hour pii audio redaction* $0.05/hour pii redaction $0.08/hour auto chapters $0.08/hour entity detection $0.08/hour content moderation $0.15/hour topic detection $0.15/hour
• Lemur: lemur default $0.015/1k tokens $0.043/1k tokens lemur claude 2.1 $0.015/1k tokens $0.043/1k tokens lemur basic $0.002/1k tokens $0.005/1k tokens

Who Is It For

• Speech recognition scientists
• Audio analysts
• Digital developers
• Technical researchers
• Data engineers