There's an irony in podcast summaries that we've thought about a lot: the whole point of a podcast is that it's audio — something you can absorb while driving, walking, cooking. But most summaries hand you a wall of text and expect you to sit down and read it.
That's not how people actually want to consume audio content.
Today we're launching audio summaries for Pro subscribers. Hit the listen button on any summary and DriftNote will generate a natural, conversational spoken recap — delivered in your choice of voice and style.
What audio summaries are (and aren't)
This is the important part: we didn't build a text-to-speech reader.
Reading a structured summary out loud is genuinely bad. You'd hear:
"Overview. This episode covers... Key Topics. Number one: the role of... Main Takeaways. Dash: the guest argues that..."
Nobody wants that. It sounds like a robot reading a document. The formatting that makes a written summary easy to skim makes it painful to listen to.
Instead, DriftNote first rewrites the summary into a spoken script — the kind of thing a friend would say if you asked them to catch you up on an episode they'd just listened to. Flowing sentences, natural transitions, no bullet points or headers. Something like:
"So basically, this episode was about... What's really interesting is how they framed the question around... The guest made a point that stuck with me: they argued that..."
Only once the script sounds like a real person talking does it get converted to audio.
The technical approach
The pipeline has two steps:
Step 1 — Script generation. The structured summary is sent to Gemini with instructions to rewrite it as spoken prose: no formatting, natural transitions, 700–800 words (about 5 minutes when read aloud), written from the perspective of someone who just listened to the episode.
Step 2 — Voice synthesis. The spoken script is converted to audio using Gemini 2.5 Flash TTS. You can pick from 8 voices — Kore, Puck, Aoede, Charon, Leda, Fenrir, Zephyr, or Vesta — and 4 delivery styles: Warm & Engaging, Professional, Casual, or Energetic.
The generated audio is cached, so the first time you listen to an episode it takes a few seconds to generate, but replaying it is instant. A persistent mini player sits at the bottom of the dashboard so you can browse your summary history while listening.
Who this is for
The obvious use case is commuting and exercise — you want the key ideas from an episode you missed, but you're not in a position to read.
It also works well as a preview. Before deciding whether to listen to a full episode, you can hear a five-minute summary and decide if the specific content is relevant enough to be worth the full 90 minutes. Think of it as a smarter version of checking the show notes.
And for people who've already listened to an episode, audio summaries are a surprisingly effective memory refresh. Hearing the main ideas spoken out loud, in a condensed form, tends to stick better than re-reading a text summary.
How to use it
Audio summaries are available on every summary in your dashboard. Look for the Listen to Summary button below any episode.
The first time you play it, allow a few seconds for the script and audio to generate. After that, it's instant. Voice and style preferences can be set in your settings page and applied to all future summaries.
Audio summaries are exclusive to DriftNote Pro ($9.99/month), which also includes unlimited summaries, Ask AI follow-up questions, and priority processing.
If you're already a Pro subscriber, the button is live in your dashboard right now. If you're on the free plan and want to try it, you can upgrade at any time — your existing summary history carries over.