A shell lesson in English turned into a Japanese podcast

A Software Carpentry lesson on grep and find, sent in English, became a two-host podcast in Japanese. Hear it, the request, its cost and our check.

How this was made. Audio and video overviews, slide decks, infographics and quizzes are made with Google's Gemini Notebook (formerly NotebookLM), on accounts that AutoContent operates. AutoContent has no agreement with Google. Google can limit or stop this access at any time.

The podcast (Japanese)

The Unix Shell: Finding Things

PodcastJapanese20 min 19 s

0:0020:19

Transcript excerpt

Speech-to-text of the output, not edited and without speaker labels. It is part of the sample output.

実はこれシェルを使いこなす上で初心者の方が最もつまずきやすい非常に重要なポイントなんです そうなんですか もし引用符を使わずにfind.-name.txtと打ってしまうと 予期せぬトラブルに巻き込まれる可能性が高いんですよ トラブルですか なぜかというとシェルにはワイルドカード つまりアスタリスクを見るとコマンドを実行する前に気を利かせて勝手にファイル名に展開してしまうという少しおせっかいな機能があるからなんです おせっかいな機能 つまりシェルが私たちが頼んでいないことまで先回りしてやっちゃうっていうことですか そうなんです例えばあなたが今採用しているフォルダの中にたまたま numbers.txt というファイルが一つだけ存在していたとしましょう はい numbers.txt ですね引用符なしでコマンドを打つとシェルは あスタリスク.txt というのはここにある numbers.txt のことですねと勝手に解釈して置き換えてしまうんです

AI-generated with AutoContent APIDownload MP3Full transcript (TXT)
Length
20 min 19 s
Credits used
10 credits
Cost on the Starter plan
€0.29
Generation time
13 min 33 s
Format
Podcast
Job
Course lessons
Language
Japanese
Generated on
2026-11-02
Make this with the API

This output used 10 credits, which is €0.29 on the Starter plan (€29 per month for 1,000 credits). Failed generations use none. Generation time runs from the request to the finished file; it depends on the sources, the options and the queue.

What came out, and what we would change

The source is in English; the request sets language to 日本語, keeps duration at default, and asks to keep the command names in English. The result is a two-host episode of 20 min 19 s in Japanese that says grep, find, ls and the file names in English, as asked.

The lesson's content is all there: grep as global regular expression print, the haiku file from a 1998 competition in Salon magazine and its "My Thesis" not found line, find with -name, why "*.txt" needs quotes so that the shell does not expand it to numbers.txt, grep with $(find ...) to search every text file, the limit to text files, and the closing quote from Alfred North Whitehead.

Two additions: a short history of grep and the memory limits of 1970s Unix, which the lesson does not tell, and a comparison of ls and find as a look around one room versus a search of the whole building, which helps. The transcript writes some English words by ear in katakana.

The episode took 13 min 33 s to generate and used 10 credits. Like our English episodes with default, it runs about 20 minutes. We did not test short in Japanese; in French it gave 5 min 29 s, and in Brazilian Portuguese 6 min 41 s.

The exact request

The body that was sent to POST /content/Create, unchanged, and the same request as curl.

request.json
{
  "resources": [
    {
      "type": "website",
      "content": "https://swcarpentry.github.io/shell-novice/07-find.html"
    }
  ],
  "outputType": "audio",
  "duration": "default",
  "text": "Teach this lesson to Japanese-speaking beginners: grep and find, with one example each, and how to combine them. Keep the command names in English. Do not mention the lesson's authors.",
  "language": "日本語"
}
curl
curl -X POST "https://api.autocontentapi.com/content/Create" \
  -H "Authorization: Bearer $AUTOCONTENT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"resources":[{"type":"website","content":"https://swcarpentry.github.io/shell-novice/07-find.html"}],"outputType":"audio","duration":"default","text":"Teach this lesson to Japanese-speaking beginners: grep and find, with one example each, and how to combine them. Keep the command names in English. Do not mention the lesson'\''s authors.","language":"日本語"}'
# then poll until "status" is 100:
curl -H "Authorization: Bearer $AUTOCONTENT_API_KEY" "https://api.autocontentapi.com/content/Status/<request_id>"

How we checked it

Transcript
We transcribed the output with Whisper large-v3-turbo and compared the transcript with the source.
Language
Detected ja, requested ja.
Loudness
-23 LUFS integrated, true peak -5.5 dBTP.
Notes
All lesson content is present and correct; the grep history is the hosts' addition.
Checked on
2026-11-02. Failed generations are run again and never published.