—
Open this page on a tablet or computer to trim and download clips. Search, playback, transcripts and saved clips still work on this phone.
Whole episode — click to listen, drag to move the selectionWhole episode — tap to listen, drag sideways to move the selection
Selection — drag to select, handles to trimSelection — drag sideways to select, red handles to trim
Loading audio
Saved in this browser, for you. Nothing here is uploaded.
Stored on this computer only. Clearing your browser data removes them.
—
Saved on this computer only. The audio stays where it is — this keeps the times.
A single zip. The audio is copied, never re-encoded.
What you're looking at, or the whole archive — your choice.
About 2 MB zipped.
What you get: one row per episode — show, title, date, length,
encore flag, tags and re-air dates, plus links to the episode page, the
original transcript, our transcript, and the audio.
Not included: transcript text itself. A single episode runs to
~49,000 characters and a spreadsheet cell tops out at 50,000 — the links
keep it readable instead of silently cut off.
Where it comes from. These rows are read from this archive — built from your two feeds and refreshed daily — not fetched from Podbean when you press the button. That is why it is instant, and why the audio and artwork are links rather than files: Podbean is still hosting those, and a spreadsheet cell cannot hold 27 MB of MP3.
| Column | What it is |
|---|---|
| show | Which programme it aired on |
| title | Episode title as published |
| published | ISO date, so a spreadsheet sorts it as a date |
| duration_min | Whole minutes, as a number you can total |
| is_encore | TRUE / FALSE — detected from the title |
| topics | Subjects, once the reading pass has been run |
| guests | Who appeared |
| interviewer | Who was asking |
| tags | People, bands, books and museums you linked in your own show notes |
| reair_dates | Every other date this programme has run |
| has_transcript | TRUE / FALSE |
| transcript_words | Word count, for judging length at a glance |
| episode_url | The Podbean page — the stable link |
| source_url | The original .srt from your feed |
| archive_url | Our cleaned transcript, readable in a browser |
| audio_url | The MP3 itself, direct from Podbean |
| guid | The feed's own identifier for the episode. Last, because it's long — but it's the one column that still means something after a rebuild, so anything referencing an episode from outside the file uses it. |
The first rows of your actual file:
Loading a sample…
Click or tap any example to run it.
Words are matched together, and the word you're still typing is matched as a prefix — so results narrow as you go, before you've finished a name.
Use an operator, quote, bracket or * and the
query runs exactly as written.
AND and OR work toodescription_textFilters combine with whatever you've searched. All clears a row; Reset clears everything at once.
Most episodes now include a full transcript, so search reaches the audio itself — not just the show notes. When a result matched something spoken, it's marked Heard in this episode and the excerpt is the actual line.
When your words were spoken aloud, the timestamps under a result show exactly where. Click or tap one and the episode starts playing from that second — no scrubbing.
Descriptions are shortened to three lines. More at the end of the last line opens the rest — the notes as published, including anything the feed carries that doesn't fit. When you've searched, the blurb shows the matching passage instead, and Show full notes opens the whole thing underneath it.
The names under each episode are people, bands, museums and books the producers linked in the show notes — a hand-curated index, not guesswork. Click or tap one to see every episode tagged with it.
Show notes say what happened, not what an episode is about. The gold-edged topics are read from the notes and the transcript, so they exist even where nobody wrote them down. Click or tap one to see everything on that subject. Unlike Tags, these are inferred — right nearly always, but worth a glance before you cite them.
Who was on, and who was asking. Read from the same pass, which catches the guests who were never hyperlinked. Click or tap any name to see every episode they appear in — the producers' links and the ones found in the audio both count.
A gold line under an episode means the same programme aired more than once — as an encore, or across programs. That's the quick answer to "have we run this already?"
Transcript beside the episode title opens the words with the tools around them, already marked up with whatever you searched for. Click or tap a line to hear it. On a tablet or computer, dragging across the text also selects a passage for the production tools.
"phrases",
AND/OR/NOT,
organiz*), with a hit count and
Enter / Shift+Enter to step
through them.On a phone this is a focused find, listen and read view. The line Edit buttons, player scissors, passage-selection lesson and download footer stay off the small screen; native text selection still lets you copy words.
Most episodes have a transcript. Those without one say so plainly rather than pretending to be broken.
Click or tap a timestamp to start playing, then press the scissors at the right-hand end of the player. That opens the audio editor on what you're listening to — the passage you clicked if you're still inside it, otherwise wherever you've scrubbed to.
The full editor needs a tablet or computer. On a phone, opening it shows that boundary plainly while search, playback, transcripts and saved clips remain available.
The clip is copied straight out of the original file without re-encoding, so it is identical in quality to the broadcast, not a second-generation copy. Cuts land within a fortieth of a second of the mark.
Two waveforms, each with its own timeline. The top one is the whole episode; the bottom one is the selection close up, and it's where the work gets done.
An edit is judged by ear, not by eye, so there are several ways to listen. The transport sits between the two waveforms:
Moves a cut point to the nearest gap between words. A cut through the middle of a syllable sounds broken; one that lands in the pause sounds deliberate. It only searches a second and a half either way, so it tidies a mark rather than moving it somewhere else.
It needs the waveform, so it stays greyed out until the audio has finished loading.
An MP3 named for where it came from — show, date, title
and timecode, like
LHPH_2026-02-12_Striking-At-Kings_4330-4353.mp3 — so a
clip sitting in a folder in six months still explains itself. The same
details are written into the file's tags.
The download button shows the size before you commit to it. All of this happens in your browser; the audio comes straight from the podcast host.
Finding a moment in the archive is work worth doing once. In the editor, ＋ Add to library opens a short dialogue where you name the clip and give it labels, then saves it — without closing the editor, so you can nudge a handle and save a second version of the same quote.
Clips in the top bar opens the list, with the count beside it.
Search clips looks through their titles, shows,
dates and labels as you type. Plain words are combined, and the last
word can be unfinished. Quotes keep an exact phrase together;
AND, OR, NOT, parentheses and
* work here just as they do in the main search.
Limit a term with title:,
show:, label: or date: — for
example label:promo NOT encore. Sort by best match while
searching, or by saved date, title and clip length. Press
/ to reach clip search; Esc clears it first.
Your own words — promo, intro, out-cue, whatever suits how you work. A clip can carry any number, and the most-used ones sit across the top of the list as filters. Choosing two shows only clips carrying both.
Typing promo when Promo already
exists reuses the one you have rather than making a second. Labels you
set carry over to the next clip you save, so a run of promos is labelled
once.
These are separate from Tags in search results — those come from the links in your show notes and are the same for everyone. Labels are yours.
Press ▶ to hear a clip without opening the editor; a scrubber appears on that row while it plays, and you can click or tap along it to move within the clip. Click or tap a title to rename it. The download arrow saves that one clip as an MP3; ⋯ holds Edit in editor, Copy link and Remove.
Removing something gives you ten seconds to undo it. Clips are grouped by when you saved them — today, yesterday, last week.
Download all packs every clip into one zip. You
can name the file and add a note — the note is saved inside as
clips.txt, alongside a list of every clip with its show,
date, times and labels, so the bag explains itself later.
If you've searched or filtered by a label, the button says
Download these 3 rather than Download all, and the file is
named …-filtered. You always know which you're getting.
In this browser, on this computer, and nowhere else. Nothing is uploaded, so the list works offline — but it doesn't follow you to another machine, and clearing your browsing data removes it. The audio itself is never at risk; only the record of what you found.
A downloaded MP3 is a normal file and outlives all of this.
The address bar always reflects what you're looking at, so any search can be copied and sent to someone — filters, sort order and all.
A link can also point at a single moment. In your Clips list, ⋯ → Copy link gives you one that opens the editor on exactly that passage. It carries a fingerprint of the episode alongside its number, so it keeps pointing at the right show even if the archive is ever rebuilt.
Export turns the archive into a spreadsheet. One row per episode, with dates a spreadsheet understands, durations as numbers you can total, and links back to the episode, its transcript and its audio.
Where the rows come from. They're read from this archive, which is built from your two feeds and checks for updates every 15 minutes — nothing is fetched from Podbean when you press the button. That's why it's instant, and why audio and artwork appear as links rather than files: Podbean is still hosting those, and a spreadsheet cell can't hold a 27 MB MP3.
Transcript text isn't in the spreadsheet, on purpose. One episode runs to about 49,000 characters and a spreadsheet cell stops at 50,000, so it would be cut off without telling you. The links get you the full text instead — and the Transcript button downloads any single episode as plain text or subtitles. The archive package below carries them as real files.
A single .zip holding the catalogue, the
transcripts and a README that explains itself — a copy of the
archive that doesn't need this website to be readable. The whole
thing, the complete catalogue and every transcript, comes to under
4 MB. It's small enough to email and plain enough that another
developer, or any AI, could rebuild the catalogue from it.
Inside:
README.md — what the package is, what every column
means, and what to know about the data. Written for someone who has
never seen this site.episodes.csv and episodes.json — one row
per episode, the same rows as the spreadsheet export.passages.csv — every spoken line with the second
it was said. This is the file for finding a
quote and knowing exactly where it is in the audio.transcripts/ — one plain-text file per episode,
named by date so the folder reads chronologically.Transcripts and timestamps are each a tick box, so you can take just the catalogue if that's all you need.
The filename says what it is. Export with
Everything selected — or with nothing filtered at all — and you
get …-complete.zip; export a filtered view and you get
…-filtered.zip, with a warning inside the README. A
partial export is a perfectly good dataset, but it isn't a backup —
and a year later nothing else would tell you which one you were
looking at.
Audio isn't in the package, only links to it.
The archive's audio runs to 8–12 GB, which is a download rather
than a file. Podbean hosts it, and audio_url points
straight at each MP3.
Anywhere
Playing an episode — with the progress bar focused
Reading a transcript
In the audio editor