# Whisper Transcriber Suite > Free, offline desktop app (BSD-3-Clause) that transcribes audio/video with OpenAI's Whisper — your audio stays on your computer, no account, no per-minute cost. Windows, macOS, Linux. Whisper Transcriber Suite runs faster-whisper, whisper.cpp or NVIDIA Parakeet locally to turn audio and video into SRT, VTT, ASS, TSV, TXT, JSON, LRC, Markdown, DOCX and PDF (plus oTranscribe, ELAN, InqScribe and Express Scribe). It adds live microphone/system-audio transcription, offline speaker diarisation with per-word timestamps, yt-dlp-powered downloads from any site it supports, adaptive denoise, a pause/resume/re-run batch queue, optional offline voice cloning (OmniVoice), optional AI summaries and a bilingual SRT pass, and a local-network mode that turns the machine into a transcription page for other devices on the LAN. Two cloud backends (a Gemini API option and Google Cloud Speech-to-Text) exist purely as opt-in extras, off by default — every default path keeps recordings on the user's own machine. Actively maintained; CI runs the full test suite on every push. Written by translation-robot. Common questions this page answers: Is it free? Yes — BSD-3-Clause, no subscription, no per-minute fee; only the ~1-3GB Whisper model itself downloads once, on first launch. Does it upload recordings anywhere? No, not by default — the two cloud backends above are opt-in and stay off until enabled. What platforms does it run on? Windows, macOS and Linux, with published Windows and macOS installers and a from-source build for Linux. How is this different from calling the Whisper API directly? No per-minute cost and no audio leaves the machine, because transcription runs locally by default instead of through a paid cloud API. ## Download - [Latest release](https://github.com/Milomilo777/whisper-transcriber-suite/releases/latest): v2.1.0 — free Windows installer, Windows portable ZIP and macOS DMG, no account or sign-up required; Linux builds from source. - [All releases](https://github.com/Milomilo777/whisper-transcriber-suite/releases): full version history, every past release kept published with its own changelog. - [Support the project — Buy Me a Coffee](https://buymeacoffee.com/mehraz1358): funds continued cross-platform development and testing. ## Guides - [Transcribe a YouTube video to text](https://whisper-transcriber-suite.com/youtube-to-text/): paste a link, download with yt-dlp, transcribe locally, save TXT, DOCX, PDF or SRT. - [Convert a video to SRT subtitles](https://whisper-transcriber-suite.com/video-to-srt/): SRT, VTT or ASS from a file or a link, correcting them, and burning them into the video. - [Transcribe with speaker labels](https://whisper-transcriber-suite.com/speaker-labels/): who said what in an interview, renaming speakers, Word and PDF export. ## Docs - [Full text for AI assistants](https://whisper-transcriber-suite.com/llms-full.txt): this summary plus the complete FAQ and README in one plain-text file. - [GitHub repository](https://github.com/Milomilo777/whisper-transcriber-suite): source code, issue tracker, discussions. - [Full README](https://github.com/Milomilo777/whisper-transcriber-suite/blob/master/README.md): who it is for, features, downloads, what leaves the computer, links to every doc. - [Install guide](https://github.com/Milomilo777/whisper-transcriber-suite/blob/master/docs/INSTALL.md): setup and troubleshooting for Windows, macOS and Linux. - [Architecture](https://github.com/Milomilo777/whisper-transcriber-suite/blob/master/docs/ARCHITECTURE.md): process model — long-lived worker processes, JSON-over-stdio protocol. - [Local-network / server mode](https://github.com/Milomilo777/whisper-transcriber-suite/blob/master/docs/SERVER.md): turns the machine into a transcription page other devices on the LAN can use. - [Configuration reference](https://github.com/Milomilo777/whisper-transcriber-suite/blob/master/docs/CONFIG.md): every config.json key, its default, and what it controls. - [Live transcription](https://github.com/Milomilo777/whisper-transcriber-suite/blob/master/docs/LIVE.md): microphone and system-audio transcription in real time. - [Build from source](https://github.com/Milomilo777/whisper-transcriber-suite/blob/master/docs/BUILD.md): the build pipelines, which two are shipped, and how to build them. - [Comparison with similar apps](https://github.com/Milomilo777/whisper-transcriber-suite/blob/master/docs/COMPARISON.md): sourced side-by-side facts for five other transcription apps. - [Multilingual benchmark](https://github.com/Milomilo777/whisper-transcriber-suite/blob/master/docs/evaluations/benchmark-v1/README.md): error rate and CPU speed of six local Whisper models in eight languages (FLEURS read speech), and the recommended model per language. - [Machine-readable facts](https://whisper-transcriber-suite.com/facts.json): version, platforms, output formats, engines, licence and links as JSON. - [Changelog](https://github.com/Milomilo777/whisper-transcriber-suite/blob/master/docs/CHANGELOG.md): what shipped in each version. ## Optional - [Adaptive denoise](https://github.com/Milomilo777/whisper-transcriber-suite/blob/master/docs/DENOISE.md): the optional pre-process that measures each recording and only cleans it when that helps. - [Cloud STT via Gemini](https://github.com/Milomilo777/whisper-transcriber-suite/blob/master/docs/CLOUD_STT.md): optional opt-in cloud backend authenticated with a Gemini API key. - [Cloud STT via Google Cloud](https://github.com/Milomilo777/whisper-transcriber-suite/blob/master/docs/CLOUD_STT_GOOGLE.md): optional opt-in Google Cloud Speech-to-Text backend with a service account. - [Design decisions](https://github.com/Milomilo777/whisper-transcriber-suite/blob/master/docs/DECISIONS.md): non-obvious design choices and why they were made. - [Roadmap](https://github.com/Milomilo777/whisper-transcriber-suite/blob/master/docs/ROADMAP.md): what is being considered next. ## Key facts - Free and open source: no subscription, no per-minute cost. - License: BSD-3-Clause for the project's own source; bundled ffmpeg, ffprobe, yt-dlp and the Whisper model keep their own upstream licenses (see THIRD_PARTY_NOTICES.md in the repository). - Offline by default: faster-whisper, whisper.cpp and NVIDIA Parakeet all transcribe locally; cloud backends are opt-in only. - Cross-platform: Windows and macOS installers published; Linux builds from source. - Local engines: faster-whisper (default), whisper.cpp, NVIDIA Parakeet. Transcription works without an internet connection after a one-time Whisper model download (about 1-3 GB); no account or API key for local use. - Output formats: srt, vtt, ass, tsv, txt, json, lrc, md, docx, pdf, plus oTranscribe, ELAN, InqScribe and Express Scribe. - Also: speaker diarisation, live microphone/system-audio transcription, a batch queue, yt-dlp downloads, and a local-network server mode. - New in 2.0: the queue shows each transcription's speed and time left; Make subtitled video turns a link into an MP4 with burned-in subtitles; a finished transcript can be saved as one offline web page; Work offline shows whether anything is connected; speech can be translated straight to English; native Mac menus, Dark mode and notifications; Windows taskbar progress and dark title bars; a quick start window with a sample clip. The Video Tiling tab was removed. - New in 2.1: the app comes in 42 interface languages (machine-translated, marked as such; Persian, Arabic and Hebrew on Windows); speaker labels reach the end of long recordings and the number of speakers can be set; files are fully written to disk before they replace the old ones; online features work on every Mac. - Latest version v2.1.0, released 2026-10-12; repository created 2026-05-11 and actively maintained. - Author: translation-robot. ## Contact - Website: https://whisper-transcriber-suite.com - GitHub: https://github.com/Milomilo777/whisper-transcriber-suite - Issues / support: https://github.com/Milomilo777/whisper-transcriber-suite/issues - Questions: https://github.com/Milomilo777/whisper-transcriber-suite/discussions - Email: shown on request in the Contact section of the website (https://whisper-transcriber-suite.com/#contact) - Security reports: GitHub private vulnerability reporting (see SECURITY.md)