Scriberr
Self-hosted audio and video transcription with Whisper, NVIDIA Parakeet and Canary. Runs offline on your hardware, with optional Ollama transcript chat.
Scriberr is a free, open source transcription app for people who want to keep meeting recordings and voice notes on their own hardware. It turns audio and video into text locally, with offline transcription once the models are downloaded. The MIT license allows you to use and modify it.
It supports Whisper, NVIDIA Parakeet and Canary, with word-level timestamps and automatic speaker detection to distinguish who spoke. The transcript reader follows audio playback, and selecting text lets you jump to the corresponding moment in a recording. You can highlight passages and add notes while listening, so reviewing a long recording doesn't require working through a separate audio player and text document.
Transcript chat connects to Ollama or OpenAI-compatible providers for summaries and questions about a recording. With a local LLM through Ollama, that processing can stay on your machine too. Choosing a cloud provider sends transcript content to that service, even though transcription itself runs locally.
Scriberr runs on macOS and Linux, or on a self-hosted server through Docker. It can use a CPU without a dedicated GPU, and supports NVIDIA GPU acceleration, including Blackwell hardware. Its browser interface also works as an installable web app on desktop and mobile. A built-in recorder captures audio, while a folder watcher automatically transcribes incoming files. An API connects transcription to automations such as n8n.
