SpeechText

SpeechText is an AI-powered speech-to-text transcription service that converts audio and video files into written text. It's useful for a variety of industries, including journalism, medicine and business, and is a good way to transcribe interviews, meetings, lectures and other audio recordings.

SpeechText uses deep neural network models to achieve a word error rate of 3.8% on the open source LibriSpeech dataset, a benchmark that's close to human performance. It works with more than 30 languages and can handle non-native speaker accents. The service also can identify speakers, so you can find out who said what in a multi-person conversation.

Some other features of SpeechText include:

• Domain-Specific Models: Models tuned for industries like finance, medicine, law and HR that can improve recognition quality.
• Audio Search Engine: Searches audio data with natural language queries.
• Automatic Punctuation: Adds commas, full stops, question marks and periods to transcriptions.
• Editing Tools: Interactive tools to proofread and verify transcription results.
• Export Options: Supports formats like txt, pdf, docx and more.

SpeechText has several pricing tiers to accommodate different needs:

• STARTER: $10 for 180 transcription minutes, 30 MB maximum file size, general models.
• PERSONAL: $19 for 380 transcription minutes, 60 MB maximum file size, domain-specific models.
• STANDARD: $49 for 990 transcription minutes, 200 MB maximum file size, domain-specific models.
• BUSINESS: $99 for 2000 transcription minutes, 1 GB maximum file size, domain-specific models.

The service also has an API for integration with applications, letting developers build speech recognition abilities into their software. The API supports several programming languages, including Python, Java, PHP and more.

SpeechText protects data with GDPR compliance and encryption of data sent from users to the service. Files and transcription results can be deleted from the dashboard at any time.

The service is designed to be easy to use, but accuracy can be affected by factors like audio quality and background noise. But with its features and pricing, SpeechText is a good option for speech-to-text transcription.

Published on June 9, 2024