WhisperPlus is an open-source Python library on GitHub that extends OpenAI’s Whisper speech recognition models with faster pipelines and extra capabilities such as speaker diarization. The project has attracted roughly 2,000 stars, is published on PyPI, and continues to receive updates, with repository activity extending into late 2026.
Whisper is powerful but bare-bones: WhisperPlus wraps it in ready-to-use pipelines so developers can go from an audio or video file to transcripts, subtitles and speaker-labelled output in a few lines of code. It builds on optimized inference backends for speed, integrates diarization models that identify who spoke when, and adds conveniences such as downloading and transcribing YouTube videos and generating subtitle files. The library exposes its features through simple Python classes, and demo spaces are available for trying it in the browser. It is a practical toolkit for anyone building transcription features into products, research pipelines or personal automation.
The project is maintained by Kadir Nar and is popular in the Turkish and international speech-tech communities alike. Its modular pipelines make it simple to swap Whisper model sizes, combine transcription with diarization, and export results for downstream tools such as video editors and LLM applications.
WhisperPlus is free and open source under the Apache-2.0 license. You only provide your own compute for running the models locally or in the cloud; the library itself has no paid plans, API keys or usage limits imposed by the project.
WhisperPlus suits developers, data scientists and creators who need reliable transcription with speaker labels, subtitle generation for video content, or speech pipelines for meeting recordings, podcasts, interviews and research datasets.