Muse Task HubUnofficial guide

Learning · Community report

Find your next podcast listen

Find your next podcast listen. Pick episodes from a large back catalogue.

What you can get

What it is

Listening shortlist

A shortlist of relevant episodes with a suggested listening order.

It came back right when

  • Every pick has an episode number, a date and a length.
  • Each reason refers to something specific in the episode rather than a general description of the show.
  • Uncertain matches are flagged rather than presented as confirmed.
  • The list is the size you asked for, in a suggested order.
  • Nothing was subscribed to, downloaded or queued.

The checks are editorial, not from the source.

Illustration — suggested task flow, not an execution result
  1. Name the criterionStops for your approval

    What you want from the catalogue, and how much time you have.

  2. Draw the shortlist

    Numbers, dates, lengths, links and specific reasons.

  3. VerifyStops for your approval

    Spot-check numbers against the show's own list.

  4. Order it

    A suggested order with dependencies stated.

Editorial, not from the source

Works when

  • The show has a large back catalogue you will never work through.
  • You can say what you want from it beyond "the good ones".
  • You will sanity-check a couple of episode numbers before starting.

Does not work when

  • You want a definitive ranking.
  • The show is small enough to browse yourself.
  • The feed is private, paywalled or hard to reach.

Find your next podcast listen: review template

A review template. Nothing is subscribed to, downloaded or queued.
EpisodeWhy it is hereChecked
[Number, date, length][Specific to this episode][Found in the show's list]
[Order][Start here, then this][Dependencies stated]
[Uncertain pick][Could not be verified][Flagged, not dropped silently]

The source case

On X, @never_offseason opens by saying he was skeptical that AI assistants do anything useful, then reports giving Muse a try over a weekend: it booked a restaurant and a parking reservation, and he asked it to research the best of 700-plus episodes of a podcast series he likes. The post is longer than the public endpoint serves, so only the opening is quoted here. It does not show the shortlist or say how the episodes were chosen, and nothing here has been run by this directory.

A public report by someone outside this library. Not independently reproduced here.

From the source

Source excerpt

I gave @Muse a try this weekend. It booked me a restaurant and parking reservation. OK whatever, thanks. I asked it to research the best of 700+ episodes of a podcast series that I like. Easy. Thanks again.

@never_offseason · Read it at the source · checked

Quoted from the original posts. Not written by this guide, and not a result produced here.

Suggested prompt

Bring Podcast, Interests, How many episodes and how much time. Acts Prepares a listening list. Does not subscribe, download or queue anything on your behalf. Source community report, checked 2026-09-18.

Written by this guide as a starting point, not quoted from the source. Your edits stay in this browser.

Other ways to ask

Editorial rewrites of the brief above, for a different emphasis. Copy them separately.

What you need

Editorial, not from the source

You supply

  • Podcast
  • Interests
  • How many episodes and how much time

Connected

  • The show's episode list or feed, read-only

Approved by you

  • No authorization to change your subscriptions or queue

How to try it

Editorial, not from the source

  1. Name the podcast and say what you want out of it: a topic, a guest, an era, or simply the ones people consider the best.
  2. Say how many episodes you want and how long you have, since a shortlist of forty is the same problem you started with.
  3. Ask for each pick to carry a reason, the episode number and date, and the length, with a link where one exists.
  4. Check two or three against the show's own episode list before you commit an evening to the order.

Say what "best" means to you

A back catalogue of several hundred episodes has no single best, and the word does most of the damage in this task. Decide what you actually want: an introduction to the show, the episodes about one subject, the interviews with people you have heard of, the ones that hold up years later, or simply the most discussed. Say how long you have, because ten hours of listening and a four-hour drive produce different lists. If the show changed hosts or format, say which era counts. Every one of these constraints removes a class of wrong answers, which is more useful than any amount of asking for high quality.

Demand checkable picks

Ask for the episode number, the publication date, the length and a link for every pick, plus a reason that refers to something specific in that episode. This is not bureaucracy: it is the only practical defence against a list that reads beautifully and contains episodes that do not exist. Shows renumber feeds, re-release old material, and publish specials outside the main sequence, so numbers drift even when the titles are right. Spot-check two or three against the show's own episode list before you start listening, and treat anything that cannot be found as unconfirmed rather than as a near miss.

Order it for the way you listen

A shortlist is more useful with a suggested order and a reason for it: start here for context, then this one for the argument against, keep this for the long drive. Ask for lengths to be shown so the list can be matched to real gaps in a week. Where an episode depends on an earlier one, that should be stated rather than left for you to discover twenty minutes in. Nothing here needs to touch your subscriptions or queue; a list you can act on yourself is both safer and easier to revise when the first pick turns out not to be your thing.

Limits and confirmation points

Limits to confirm

  • Episode titles and numbers are a common place for confident errors, especially for shows with renumbered feeds, specials or re-releases.
  • "Best" is a judgment. Unless you name a criterion, the shortlist reflects whatever the assistant weighted, which may be popularity, recency, or nothing you would recognize.
  • Back catalogues are uneven: hosts change, formats change, and an episode praised for its guest may be unrepresentative of the show.
  • This is one attributed report. It does not establish that the picks were good or the summaries accurate.

Stop and check

Editorial, not from the source

  • Before it starts: state the criterion and the size of the list.
  • On delivery: check two or three episode numbers against the show's own list.
  • Before a long drive: confirm the episodes actually exist in your app.

If it goes wrong

Editorial, not from the source

An episode number does not exist
Ask for the picks to be re-derived from the show's own episode list, and treat any pick without a verifiable number as unconfirmed.
The list is all recent episodes
Say so explicitly: cap how many can come from the last year, or name the era you want covered.
Summaries sound generic
Ask what specifically happens in each episode; a summary that would fit any episode of the show is a sign the episode was not actually examined.

Common questions

Why insist on episode numbers and dates?

Because they are checkable. Titles are easy to approximate and shows renumber their feeds, so a number and a date let you confirm the episode exists before you plan an evening around it.

What should a reason for a pick look like?

Something specific to that episode: the guest, the argument, the thing that happens. A reason that would fit any episode of the show is a sign the episode was not really examined.

Can it just rank the whole catalogue?

It can produce a ranking, but the ranking will reflect an unstated criterion. Naming the criterion yourself is what makes the result usable rather than merely confident.

Will it add episodes to my app?

Not as written here. It prepares a list; subscribing, downloading and queueing stay with you, which also makes it easy to discard picks that miss.

How many episodes should I ask for?

Few enough to start today. Five with an order beats forty ranked, and you can always ask for the next five once you know whether the criterion was right.

Next in Learning

Get recurring task recommendations — ask your own agent to check for relevant updates.