Internet Archive
Preserving AI audio history.
Why We Archive
My Weird Prompts is one of the first podcasts produced almost entirely by artificial intelligence — from script generation to voice synthesis to cover art. Every episode begins as a short voice memo from a human and emerges as a full-length podcast episode created by an automated pipeline.
We believe this makes the show worth preserving. Not because every episode is a masterpiece, but because together they form a unique record of what AI-generated audio sounded like at this moment in time. The voices, the pacing, the occasional odd phrasing, the way two AI hosts riff on a human prompt — all of it is a snapshot of a technology that's evolving fast.
The Internet Archive is the natural home for this kind of preservation. Their mission — universal access to all knowledge — aligns with our belief that this experiment should be openly accessible, not locked behind a single platform. If Spotify or Apple Podcasts disappeared tomorrow, the archive would still be there.
What's Preserved
Each episode is uploaded as an individual item on the Internet Archive, containing:
- The full episode audio, converted to MP3 for maximum compatibility
- A complete transcript of the episode
- The AI-generated cover art
- Full metadata — title, date, tags, season, and category
Everything is free to stream, download, and redistribute under a Creative Commons BY-NC 4.0 license. No account needed, no paywall, no restrictions on access.
An Experiment at Scale
With hundreds of episodes and counting, MWP is among the largest collections of AI-generated podcast content in existence. The archive captures the evolution of the show's production pipeline — from early episodes with simpler scripts and rougher voices to later episodes with multi-pass editing, grounded fact-checking, and more natural-sounding speech synthesis.
For researchers, journalists, or anyone curious about the trajectory of AI audio, this collection offers a longitudinal dataset that would be difficult to reconstruct after the fact. We'd rather archive it now while the context is fresh than wish we had later.