Thirty years old, free on every desktop platform, and more capable than most paid editors. It also looks its age, and that is the whole story.
If you ask ten people what to record with and they are not being paid to say otherwise, roughly eight will say Audacity. It has been the default free audio program since the early 2000s, it runs on Windows, macOS and Linux, and in that time it has accumulated a feature list that embarrasses software costing hundreds. It is also, visually, from another era, and for a lot of people that is a feature, because nothing moves around between versions.
The installer is small and the program opens in a couple of seconds to a mostly empty grey window with a toolbar across the top. Pick your input device from the dropdown, press the red button, and you are recording. That part is genuinely as easy as a browser recorder.
What trips new users is the second session. Audacity saves its work in a project file (.aup3), not an audio file. Pressing Save gives you a project, not something you can email. To get an MP3 or WAV you use File → Export, which is one of the most-searched things about the program for good reason. Once you know, you never forget.
Audacity records as many tracks as your interface can feed it, with punch-in, overdubbing against existing tracks, and a timer-record feature that starts and stops unattended, which is useful for capturing a radio broadcast or a scheduled stream. It writes to disk as it goes, so a crash costs you seconds rather than the session, which is the single biggest advantage it holds over any browser-based recorder.
On Windows it can also record system audio without any extra software: switch the audio host to WASAPI and choose the loopback version of your output device. That one trick covers most of what people install screen recorders for. On macOS and Linux you still need a loopback driver such as BlackHole or PulseAudio's monitor source.
This is where Audacity earns its reputation. The tools that matter for spoken word are all present and all free:
Since version 3.2 the main effects can run in real time from a per-track effects rack, so you can hear a compressor working while the audio plays instead of applying, listening, undoing, and applying again. That change alone modernised the program more than any redesign would have. It also loads VST3, LV2 and Audio Unit plugins, so the free plugin world is available to you.
Audacity is a destructive editor at heart. Effects are baked into the audio when applied, and while undo history is long, it is not the non-destructive, always-reversible model that a modern DAW uses. If you want to change a compressor setting from three edits ago, you are re-doing three edits.
Latency handling is workmanlike rather than good. Recording a vocal while monitoring a backing track through the software needs manual latency compensation, and getting it right is fiddly. Musicians overdubbing regularly will be happier in REAPER or GarageBand.
And the interface, while consistent, is dense. Everything is in a menu, the menus are long, and discovering features happens mostly by searching the internet rather than by exploring the program.
Audacity was acquired by Muse Group in 2021, and a proposed telemetry addition caused a genuine uproar that ended with the plan withdrawn and the privacy policy rewritten. The program remains open source under the GPL, development has actually accelerated since, and forks exist for anyone who remains uncomfortable. Worth knowing about; not, in our view, a reason to avoid it.
The correct first install for almost anyone who records more than occasionally. It costs nothing, it does the boring important things properly, and the ceiling is far higher than the plain interface suggests. Learn Export, learn Noise Reduction, and you have covered most of what paid editors sell.
Move on when you need non-destructive multitrack work with live monitoring, which is REAPER's job, or when you need spectral repair, which means Audition.