Loading...
Loading...
Updates, engineering deep dives, and thoughts on the future of podcast search.
Today we're launching PodSearch.io, the first search engine built specifically for finding content inside podcast episodes using AI transcription.
A deep dive into our transcription pipeline: from audio ingestion to searchable text. Learn how we use Whisper and custom models to achieve high accuracy at scale.
Why we chose PostgreSQL tsvector over Elasticsearch for our search infrastructure, and how we handle millions of transcript segments with sub-second query times.
Over 4 million podcasts exist, but finding specific conversations is nearly impossible. We're changing that with transcript-level search.
We tested our transcription pipeline against industry benchmarks. Here's how PodSearch.io performs across different languages, accents, and audio conditions.
We're rolling out user accounts with the ability to save episodes, follow your favorite podcasts, and keep track of your search history.
Get notified when we publish new articles about podcast search, AI transcription, and product updates.