Rev AI
Rev AI is a speech recognition and transcription API platform from Rev.com that delivers high-accuracy AI-powered speech-to-text for pre-recorded audio and real-time streaming. The platform supports asynchronous batch transcription, real-time streaming via WebSocket, topic extraction, sentiment analysis, language identification, forced alignment, and custom vocabulary to improve accuracy. Rev AI uses Bearer token authentication and offers pay-as-you-go pricing starting at $0.10/hour for Reverb Turbo transcription, plus a free tier of 5 hours of credits for new accounts. Official SDKs are available for Python, Node.js, and Java through the revdotcom GitHub organization.
APIs
Rev AI Asynchronous Speech-to-Text API
Batch transcription API for pre-recorded audio files supporting uploads up to 2 GB (multipart) or 5 TB (source URL), with results typically available within 15 minutes.
Rev AI Streaming Speech-to-Text API
Real-time speech transcription via WebSocket connections supporting up to 10 concurrent streams with a 3-hour time limit per stream.
Rev AI Topic Extraction API
Extracts key topics and insights from transcripts generated by the Rev AI speech-to-text pipeline.
Rev AI Sentiment Analysis API
Provides sentiment insights for transcripts, priced at $0.0008 per 10 words analyzed.
Rev AI Language Identification API
Detects the spoken language of audio input, priced at $0.003 per minute.
Rev AI Forced Alignment API
Aligns existing transcripts with audio for precise word-level timestamps, priced at $0.003 per minute.
Rev AI Custom Vocabulary API
Allows submission of custom vocabulary lists to improve transcription accuracy for domain-specific terminology.