What is in 4a's library

A picker is only as good as what it has to pick from, so this page answers the two questions people actually have: why is this show in here, and why is that one not.

The shape of it

4a draws on two pools with different jobs.

A curated set. A few hundred shows that somebody judged — not by chart position, but by whether the show is good: whether it says something, whether the episodes are substantive rather than trailers and re-runs. This is the pool the daily cards mostly come from.

A much larger breadth set. Show-level records across essentially every subject category the public podcast directories carry, gathered from public charts across a long list of countries. This is what stops the library from being a list of the same forty famous shows, and it is where the Stretch queue goes hunting.

Measured, 2026-08-25: the curated pool covers 220 shows, and the breadth pool holds records for tens of thousands more. A show moves from the second pool to the first by being judged, not by getting more popular.

Where the range is wide

The subjects that do well in 4a are the ones where a queue makes more sense than a single episode: engineering, history, craft and making things, science, business, the long-interview shows, and the sort of specialist show that has forty hours on one narrow thing and no obvious way in.

That last category is the one 4a is most useful for and the one a chart is worst at. A show with a small, devoted audience and a deep back catalogue is invisible on a popularity list and is exactly what you want on a Tuesday.

Being obscure is not a disqualification

Worth stating flatly, because most recommendation systems quietly do the opposite.

Chart position is used as a hint about quality when we know nothing else about a subject — that is all. It is not a gate. A show is not excluded for being small, for being independently hosted, or for having no marketing behind it.

The same applies in the other direction: a show is not included merely because it is enormous.

What is not in there

• Anything without a public feed. 4a works from the feeds publishers publish. If a show is only inside one app, or is subscriber-only, 4a cannot see it and does not try to get around that.
• Video-first shows, in general. 4a is for listening.
• Anything a publisher has asked us to remove. For podcasters is the page for that, and the answer is yes.
• Everything. The library is not the whole of podcasting and does not pretend to be. If your favourite show is missing, tell us — that is a useful email and it gets read.

The subject labels are imperfect, and here is the honest reason

Every show in the library carries subject labels, and those labels are what put it in one queue rather than another. Two things make them wrong sometimes.

The first is us. Some labels are assigned automatically, and automatic classification of a podcast from its description is genuinely hard. A general-audience science show can end up filed under medicine, confidently. When we find one of those we fix it, and there is a standing rule that something doubtful gets labelled rather than silently dropped — because a show quietly excluded is a show nobody ever notices is missing.

The second is not us, and cannot be fixed by trying harder. A large part of a show's subject signal comes from the category the publisher chose to list themselves in, and publishers choose that for their own reasons, some of which are marketing rather than description. A paranormal storytelling show listed by its publisher under Natural Sciences will arrive carrying science. We can hedge against our own classifier being wrong; we cannot hedge against a publisher filing themselves somewhere surprising.

So: a subject label is a strong hint and not a promise. If a queue contains something that clearly does not belong, that is the mechanism above, and telling us is the fastest way it gets fixed.

How a show ends up in the daily cards

Three things have to be true.

1. It is in the library — see above.
2. It matches a subject you have shown an interest in, or, for the fourth card, sits next door to one.
3. There is something recent and substantive to actually play. A subject with nothing good in it this week does not get a card. 4a would rather give you three solid queues and one narrow one than four padded ones.

How the library stays current

New episodes arrive because publishers publish them and 4a asks their feeds whether anything has changed. Asking politely matters more than asking often: a podcast host serving hundreds of shows should not be able to tell that we are looking. That is a real constraint on how fast a brand-new episode shows up, and we would rather be a little late than be a nuisance to someone whose bandwidth we are spending.

Where audio comes from

Always the publisher's own servers. 4a never rehosts, proxies, transforms or re-encodes podcast audio, and there is no 4a server in the path between you and an episode. Nothing is downloaded to us, nothing is cached by us, and nothing is modified.

We also do not detect, skip or strip advertising. Not as a favour — as a commitment that shapes the entire product. When you play something, you get the publisher's file as the publisher serves it.

Suggest a show

If something is missing, send it to help@jwlabs.ai. A feed URL and one line about why it is good is enough. We do not promise to add everything, and we will say so rather than going quiet.

Publishers: there is a page for you.