AssemblyAI has introduced Universal-1, a new speech recognition model that offers near-human accuracy for automated speech recognition (ASR). It is designed to transcribe audio files with accents, background noise, and complex phrases. The model is accessible through the same web API as previous ASR models.
To help developers, AssemblyAI offers an official Python SDK for easy installation, requiring a signed-up account for an API key. The SDK includes two pricing tiers: Best for maximum accuracy and Nano, a cost-effective solution supporting transcription in 99 languages. By default, the SDK uses the Best tier.
To switch to the Nano tier, developers can adjust the TranscriptionConfig object. Beyond transcription, AssemblyAI offers advanced features like entity detection, content moderation, PII redaction, and Large Language Model (LLM) applications for audio data. These features make the transcription service suitable for a wide range of applications.
Source
<p>The post Transcribe Audio Files with AssemblyAI’s Universal-1: Near-Human Accuracy & Affordable Pricing first appeared on CoinBuzzFeed.</p>