Audio overviews
How to Turn Your Substack Archive Into a Podcast Without Recording
You have written hundreds of thousands of words. Here is how to let listeners hear them, with every claim traced back to the post it came from.
Why most AI podcasts from blogs feel off
Generic text to speech tools read your post out loud. Generic AI podcast tools go the other way and improvise, adding facts you never wrote and opinions you do not hold.
Readers notice both. The first sounds like a robot reading a newsletter. The second puts words in your mouth.
What a grounded audio overview does differently
Notestack first indexes your posts into passages. The script writer can only use those passages, and every line it writes carries a pointer to the passage that supports it.
- Two hosts with distinct voices discuss your ideas, not the internet's
- Every factual line maps to a cited passage you can click
- Unsupported statements are rejected before audio is generated
If a host says it, you wrote it.
The workflow
Paste your Substack URL. Pick a post or build a notebook of related posts. Choose a format: deep dive, brief or debate. Notestack writes the script, you review it, and ElevenLabs voices it.
Publish the file to your podcast feed, attach it to the original post, or cut a 60 second audiogram for social.
Want to see the shape of the episode before you connect anything? Paste a post into the free AI podcast generator for a two host script, or use the podcast script generator if you plan to record it yourself.
Questions
Do I need to record my own voice?
No. Notestack ships with a curated pair of host voices. Voice cloning is optional and only happens after you record an explicit consent sample.
Can the hosts make things up?
The script generator only sees passages from your posts, and a separate judge checks each factual line against its citation. Lines that fail are rewritten or removed.