Some of the most revealing product decisions are very small. They do not headline a release, they do not generate a lot of screenshots, and they are usually invisible unless the software gets them wrong. The recording metadata sheet in Kanora falls squarely into that category.
On the surface it is just a post-capture form. Title, artist, album, source, a little file information, an import button. Nothing about that sounds especially charged. But the moment I started shaping the live-audio workflow, it became obvious that this sheet was going to carry a lot more of the app’s attitude than its size would suggest. If the recording flow treated every source as a generic anonymous upload, the rest of the analog story would feel fake no matter how polished the monitoring and capture path became.
That is why the source field matters, and why it was worth making the app a little more aware of what people are likely to type there. If someone says the source is vinyl, record, tape, cassette, radio, or stream, the software should recognise the shape of that input and respond accordingly. Not with fireworks, and not by forcing them through a novelty workflow, but simply by behaving like these are normal, legitimate things to preserve. Even something as small as offering a sensible genre suggestion or understanding the semantic difference between a tape transfer and a radio capture changes the tone of the feature.
I keep coming back to tone in product work because it is often the difference between software that merely contains a feature and software that seems to understand why the feature exists. A lot of modern media tools are built around upload metaphors that flatten everything into the same shape. That works fine if the only thing you care about is moving files around. It works much less well if you are building for people whose collections include objects with provenance, history, and odd little contextual details that matter precisely because they are not interchangeable.
This is one of the places where Kanora’s library-first approach has been useful. The recording sheet is not trying to be a social capture flow or a content funnel. It is a bridge between the moment of recording and the longer life of the object in the library. That changes the priorities. File details are worth surfacing. The source matters. The metadata does not need to be exhaustive, but it does need to be enough to stop the recording from becoming another orphaned file that only makes sense on the day it was made.
AI’s role here has been fairly mundane, which is probably appropriate. It is helpful for quickly enumerating edge cases in form behaviour, checking the shape of a flow, or suggesting where a little more context would reduce friction. It is not especially useful for inventing product sympathy. That part still has to come from knowing the kind of user you are building for and paying attention to what feels condescending, too vague, or too eager to turn a niche workflow into an “experience.”
I think that is why I like this part of the app more than I expected. It is modest, but it says the right things quietly. Vinyl and cassette are not treated as cosplay objects. Radio and live sources are not forced into the same mould as a clean file import. The software simply makes room for them and assumes they belong. For a project like Kanora, that kind of assumption matters. It is a way of telling the user that their collection is allowed to have texture.