CLI app to get podcast feed URLs

dugan · 06-24-2019, 09:46 PM

I've been getting into podcasts recently, and I've noticed that, a) a lot of clients expect you to just give them the RSS feeds that you want to subscribe to, and b) those RSS feeds aren't always easy to find.

So I kludged up a small script to query for them.

Code:

#!/usr/bin/env bash

if [[ "$1" == "" ]]; then
  exit 1
fi

PODCAST=$(http "https://itunes.apple.com/search?media=podcast&term=$1" | jq -r '.results[].collectionName' | fzf --select-1 --exit-0)

if [[ "$PODCAST" == "" ]]; then
  exit
fi

echo "$PODCAST"
http "https://itunes.apple.com/search?media=podcast&term=$PODCAST" | jq -r '.results[0].feedUrl'

Name it "podfeed" and call it like this:

Code:

podfeed "You Must Remember"

That currently gives you a menu of two choices:

You Must Remember Manson
You Must Remember This

Narrow down the list by typing, or use the arrow keys to make a selection, and press ENTER.

Then you'll see:

Code:

You Must Remember Manson
http://feeds.megaphone.fm/PPY3689373157

Note that this uses a fairly modern pipeline, specifically:

I'll decline to link to the iTunes search API documentation, as all I did was look at what CPod does:

https://github.com/z-------------/CP...ch-podcasts.js

For more ideas on what you can do with FZF, see here:

How FZF and ripgrep improved my workflow

individual · 06-25-2019, 07:50 AM

That's pretty neat, Dugan. My only question is whether it would be better to store the result of the first search in a variable, that way you don't need to call the same URL again?
Something like this?

Code:

#!/usr/bin/env bash

data="$(<1.txt)"
feedNames="$(<<<"$data" jq '.results[].collectionName')"
<<<"$data" jq -r ".results[] | select(.collectionName == $(fzy <<< "$feedNames")) | .feedUrl"

EDIT: By the way, I use fzy, but it works just the same with fzf.

dugan · 06-27-2022, 06:32 PM

I wrote another implementation using skim. This one will just give you a prompt and an autocomplete. It's slightly less responsive because it's sending a request to itunes with each keystroke.

pdfd script to make it work:

Code:

#!/usr/bin/env bash

if [[ "$1" == "" ]]; then
  exit 1
fi

http "https://itunes.apple.com/search?media=podcast&term=$1" | jq -r '.results[].collectionName'

The new podfeed script you actually run:

Code:

#!/usr/bin/env bash

PODCAST=$(sk --ansi -i -c 'pdfd "{}" &')

if [[ "$PODCAST" == "" ]]; then
  exit
fi

URL=$(http "https://itunes.apple.com/search?media=podcast&term=$PODCAST" | jq -r '.results[0].feedUrl')

echo "$URL" "$PODCAST"

And do you want to know what else I did? I wrote my own podcast aggregator, Podfeeds.

dugan · 04-29-2026, 09:03 PM

I just set up a new Fedora 44 installation, and the Skim version didn't work! I don't know if it was a breaking change to Skim or what.

Anyway, I just made a slight change to the top version. It now formats its output in a way the way my podcast client wants:

Code:

#!/usr/bin/env bash

if [[ "$1" == "" ]]; then
	exit 1
fi

PODCAST=$(http "https://itunes.apple.com/search?media=podcast&term=$1" | jq -r '.results[].collectionName' | fzf --select-1 --exit-0)

if [[ "$PODCAST" == "" ]]; then
	exit
fi

URL=$(http "https://itunes.apple.com/search?media=podcast&term=$PODCAST" | jq -r '.results[0].feedUrl')

echo - "$URL" \# "$PODCAST"

(As we've seen, some people are better at BASH than me; fork as needed).

dugan · 05-23-2026, 06:33 PM

When I saw https://junegunn.github.io/fzf/tips/...p-integration/ I knew it was the right way to go. I finally got around to rewriting these scripts that way.

The helper _podfeed_query script:

Code:

#!/usr/bin/env bash

if [[ "$PODFEED_JSON" == "" ]]; then
	echo Please set PODFEED_JSON
	exit 1
fi

if [[ ! -f "$PODFEED_JSON" ]]; then
	echo PODFEED_JSON file not found
	exit 2
fi

# See: https://performance-partners.apple.com/search-api
curl --silent \
	--get \
	--data-urlencode "media=podcast" \
	--data-urlencode "entity=podcast" \
	--data-urlencode "attribute=titleTerm" \
	--data-urlencode "term=$1" \
	https://itunes.apple.com/search >"$PODFEED_JSON"

cat "$PODFEED_JSON" | jq -r '.results[].collectionName' | sort | uniq

The podfeed script you actually run:

Code:

#!/usr/bin/env bash

# Based on this tutorial:
# https://junegunn.github.io/fzf/tips/ripgrep-integration/

PODFEED_JSON=$(mktemp)
export PODFEED_JSON

# I couldn't get the start: binding to work.
# Fedora 44. FZF 0.70.0.

_podfeed_query "$1" | fzf \
	--disabled \
	--ansi \
	--select-1 \
	--bind "change:reload:_podfeed_query {q}" \
	--bind "enter:become:cat $PODFEED_JSON | jq -r '.results[0] | (\"- \" + .feedUrl + \" # \" + .collectionName)'"

rm -rf "$PODFEED_JSON"

Note I've dropped the HTTPie dependency. It's also much more efficient. Doesn't really matter, but it is. Yes, the inefficiency that was pointed out earlier has been addressed.

This will be the last time I update this thread, as the scripts now have a home here:

https://github.com/duganchen/podfeeds/tree/main/scripts