whisper.cpp Server

Overview

whisper.cpp is a high performance implementation of OpenAI's Whisper automatic speech recognition model written in plain C and C++. It runs entirely on the CPU with no GPU, no Python runtime and no external services, which makes accurate speech to text practical on ordinary general purpose hardware.

It ships a bundled HTTP server that turns the engine into a service: post an audio file and receive a transcript back as plain text, JSON, or SRT and VTT subtitles. A browser page lets you upload a file and read the transcript without writing any code. Many audio and video formats are accepted and converted automatically, so recordings, meetings, voice notes, interviews and podcast episodes can be transcribed as they are.

Because everything runs locally, audio never leaves the machine you control, which suits recordings that cannot be sent to a third party transcription service.

Common uses

• Self hosted transcription of meetings, interviews and voice notes
• A private speech to text API for audio that cannot leave your control
• Generating SRT and VTT subtitles for recorded video

See it running

Real screenshots taken while testing this image against its deployment guide.