Taking the audio out of a video
A video file is a container holding separate video and audio streams alongside each other. Extraction discards the video stream and keeps the audio, which means there is no quality loss involved in the extraction itself — you are copying out a stream that was already there, not converting a picture into a sound.
MP4, MOV, MKV, AVI and WebM all work, along with most other common containers. Which audio codec sits inside varies — usually AAC in MP4 and MOV, often Vorbis or Opus in WebM — but that is handled for you; you choose the output format and the extraction takes care of the rest.
Worth being explicit about, because it differs from the audio tools. Reading a video container and demuxing its streams needs a real decoder, so the file is sent over an encrypted HTTPS connection to be processed, then deleted as soon as your download completes. Nothing is stored, nothing is inspected, and no copy is kept — but the upload does happen, and video files are large, so expect the transfer itself to take the bulk of the time.
The 500 MB limit applies to the video, not the audio you get out of it. A long recording in a high bitrate can exceed it easily. Trimming the video in another tool first, or extracting from a lower-resolution copy, is the way around it — the audio stream is usually identical either way.
Choosing the output
For lectures, interviews and anything speech-based, MP3 is the sensible target: small, universally playable, and speech survives compression well. For music extracted from a performance video, or anything heading into further editing, choose WAV or FLAC. The audio inside a video is usually already lossy, so re-encoding it to another lossy format compounds the loss — going to FLAC avoids adding a second generation of it, even though the source was not lossless to begin with.
Common uses are pulling the audio from a lecture recording to listen to while travelling, lifting a soundtrack from a performance video, saving a voice memo that was captured as video by accident, and getting a transcribable audio file out of a meeting recording. Once extracted, the Noise Remover and Audio Compressor are the usual next steps for speech.