What is WAAS?
WAAS (Whisper as a Service) is a powerful GUI and API solution that brings OpenAI's Whisper speech recognition model to your fingertips with asynchronous job queuing. It enables users to upload audio or video files for transcription, receiving results via email with download links for Jojo-files, SRT, or plain text. The Jojo-file format allows seamless editing in a local browser-based editor, where you can listen to segments and correct transcription errors by holding the Control key to play specific parts. The API provides endpoints for submitting transcription jobs, detecting language, downloading results, checking job status, and monitoring queue length. It supports webhooks with signature verification for automated workflows. WAAS is ideal for content creators, journalists, researchers, and developers needing accurate, scalable transcription services. It leverages Whisper's multilingual capabilities and multiple model sizes (from tiny to large), with VRAM requirements starting at 1GB for the tiny model. The system is designed for efficiency, processing jobs asynchronously to handle high volumes without blocking. Use cases include transcribing interviews, podcasts, lectures, meetings, and video content for subtitles or analysis. Technical details include JSON-based API responses, customizable output formats, and integration with existing pipelines via webhooks. WAAS is open-source under the Apache 2.0 license, hosted on GitHub, and built with Python and Flask. It requires a GPU for optimal performance but can run on CPU with slower speeds. The project is maintained by Schibsted and actively developed, with documentation and test examples available in the repository. Overall, WAAS simplifies access to state-of-the-art speech recognition, combining a user-friendly GUI for manual transcription with a robust API for automated processing.
Use Cases
Key Features
asynchronous job queuing audio/video transcription email delivery with links Jojo file editing webhook support language detection multiple model sizes API endpointsWho is it for?
content creators, journalists, researchers, developers, podcasters, lecturers, transcription services
Similar Tools
Details
TalkToText is an AI-powered transcription tool that converts voice to polished text instantly. Perfe...
Freemium
Details
Turbo Transcription AI delivers lightning-fast, AI-powered transcription with an impressive 99% accu...
freemium
Details
VideoToTextAI is a cutting-edge AI tool designed to transcribe audio from videos and other sources e...
Freemium
Details
Happy Scribe is an AI-powered transcription and meeting-notes platform that transforms audio and vid...
Subscription
Details
Conformer2 is a state-of-the-art speech recognition model from AssemblyAI, trained on 1.1 million ho...
Freemium
