> ## Documentation Index
> Fetch the complete documentation index at: https://koe-docs.umida.co/llms.txt
> Use this file to discover all available pages before exploring further.

# Getting Started

After downloading Koe, you need to choose a provider to do the actual speech recognition. Koe supports the following providers:

* Whisper (default), which runs locally on your computer
* OpenAI, using OpenAI's whisper API.
* Deepgram, using Deepgram's API

By default, Koe uses Whisper, which is free and runs locally on your computer. However, it is not as fast as cloud services like OpenAI and Deepgram.

To use Whisper, first you need to head to settings and download a Whisper model:

<video muted loop playsinline controls className="w-full" src="https://cdn.jsdelivr.net/gh/egoist-bot/images@main/uPic/rR8ibB.mp4" />

Generally the large the model is the more accurate it is, but it also takes more time to transcribe.

When a model is downloaded, you can start transcribing by dragging an audio file or a video file into the window:

<video muted loop playsinline controls className="w-full" src="https://cdn.jsdelivr.net/gh/egoist-bot/images@main/uPic/gOXeKL.mp4" />