Audio summarizer
AI audio summarizer: key points from a podcast, voice memo or recording
This AI audio summarizer starts from a file, not a link: upload an MP3, M4A, WAV or OGG recording, Whisper writes out the speech, and AI returns 5–7 key points or full notes of a podcast episode, voice memo, interview or meeting, with your first recording free up to 10 minutes. After that, a file uses one minute of balance per minute of audio, from $0.033. It is the sound-only branch of the AI video summarizer, and Spotify or Apple Podcasts links are not accepted.
Updated
On this page
Summarize an audio file in four steps
An audio file becomes a summary in four steps, and nothing is charged before the price is on screen.
- Choose the file. Drop an MP3, M4A, WAV, FLAC, OGG or another audio file on the upload area. Your browser checks the type and size before anything is sent.
- Start the transcription. A guest’s first recording starts straight away and free: in full up to 10 minutes, the first 10 minutes of a longer one. After sign-in the page shows how many minutes the file will use before you press Transcribe.
- Press Summary. When the transcript is ready, the Summary tile gives 5–7 key points; Detailed notes gives every topic with the facts and names under it.
- Keep the result. Copy it, or download it as PDF or DOCX; the transcript itself saves as TXT, Markdown, SRT or VTT.
Without an account the upload limit is 100 MB; after sign-in it is 500 MB and 4 hours. Every accepted format and the transcription itself are covered in the guide to transcribing audio to text.
Summarizing a podcast episode
A podcast summarizer here needs the episode as a file, because streaming apps keep downloaded episodes inside the app.
- From the show’s website or feed. Most shows publish each episode as an MP3; the RSS feed links to it directly. Download it and drop it on the page.
- From YouTube. Many shows post full video episodes there. Paste that link into the video summarizer instead: with captions, the summary costs nothing.
- From your own show. Upload the master file, then use Article for a draft of show notes and Social post for three promotional drafts.
For a podcast summary, a 60-minute episode uses 60 minutes of balance, and every AI action on it is included after that: a summary, notes, an article and a chat where you ask, for example, which books the guest recommended.
Voice memos, interviews and meeting recordings
Short personal recordings are where a summary saves the most rereading, because people speak in loops when they think aloud: summarizing a ten-minute voice memo often leaves five lines.
- iPhone Voice Memos: Share → Save to Files gives an M4A you can upload; see the guide to M4A files.
- Android recorders save M4A or MP3, ready as they are.
- Voice messages saved from Telegram or WhatsApp come as OGG or Opus, which the uploader accepts.
- Zoom audio-only recordings arrive as M4A.
A custom prompt fits these recordings well: “list every task I mentioned”, “what did we decide about the budget”, “write the reply I dictated as an email”. For a recorded class, the page on summarizing a recorded class goes into long sessions and course series.
What comes back from an audio summary
An audio summary is text: a bulleted list of 5 to 7 sentences, each carrying one point from the recording.
| Tile | Result |
|---|---|
| Summary | 5–7 key points |
| Detailed notes | every topic with facts, figures and names; see what such notes look like |
| Article | a blog post or show notes draft with 3–5 sections |
| Social post | drafts for X, LinkedIn and Instagram |
| Chat and custom prompt | answers drawn from the transcript only |
| Pro Transcript | the full text in paragraphs, filler words removed |
Results come in the language of the recording; to get an English summary of a Spanish interview, switch the transcript to English first.
What’s free and what uses minutes
The first recording is free without an account; after that, summarizing an audio recording costs one minute of balance per minute of audio, with every AI action included.
- First recording: in full up to 10 minutes, the first 10 minutes of a longer one; on a shared network, such as an office, there is a daily limit per network.
- Sign-up gift: 10 minutes of transcription, once per account, expiring after 7 days.
- Packs: 1 h for $2.99 · 3 h for $7.99 · 10 h for $19.99 · 30 h for $58.99; one-time payment, no subscription, refunds within 14 days.
The original file is deleted from the server once its sound is extracted, and the sound goes to OpenAI’s Whisper for recognition; the privacy policy has the details.
What the audio summarizer doesn’t do
The audio summarizer takes files only, so a link from Spotify, Apple Podcasts, SoundCloud or Google Drive has to become a downloaded file first.
- No speaker labels in a conversation.
- No spoken summary; the result is text.
- No recording from a microphone on the site.
- Songs and music beds summarize poorly, since lyrics carry little structure.
- Files over 500 MB or longer than 4 hours must be split first.
Everything AI does with the same transcript, from an article to a chat, is laid out in the overview of AI summaries for video and audio.
Frequently asked questions
Can I summarize audio for free?
Yes, your first recording: without an account it is transcribed free, in full up to 10 minutes and the first 10 minutes of a longer one, and its summary comes with it. After that, a file uses one minute of balance per minute of audio, from $0.033.
Do I need an account?
No for the first recording. Whole files after that need sign-in, which works through a link sent to your email, with no password.
Can I summarize a podcast from Spotify or Apple Podcasts?
No, not by link. Download the episode as MP3 from the show’s website or feed and upload the file, or paste the YouTube link if the show posts full episodes there.
Is my audio file stored?
No. The original is deleted from the server as soon as the sound is extracted, and the audio pieces are deleted after recognition. The text stays in your browser, or in your account after sign-in.
Can it tell who said what?
No. There are no speaker labels: the transcript and the summary treat a two-person interview as one stream of speech.
Does it give an audio summary I can listen to?
No. The summary is text only; there is no voice-over or spoken version.