Tapesearch

Find what was said in a podcast instantly.

Freemium

Description

Tapesearch provides a platform to explore and analyze spoken content within podcasts. It leverages AI to offer instant searching across millions of podcast transcripts, allowing users to pinpoint exact moments and discussions without listening to entire episodes. The service facilitates efficient research by enabling users to find key information, track brand mentions or specific topics with alerts, and even interact conversationally with podcast content to extract insights.

Designed for various professionals and enthusiasts, Tapesearch helps unlock a previously opaque source of information found in audio conversations. It aids in market intelligence, content creation, academic studies, advertising analysis, and journalism by providing access to timestamped transcripts, trend visualization, and API integration for developers looking to incorporate podcast data into their applications.

Key Features

Use Cases

Boost podcast content creation and discoverability
Identify market opportunities and track sentiment
Incorporate audio transcript data into academic research
Optimize advertising by tracking sponsored slots and ad volume
Spot consumer trends and track brand discussions
Find expert commentary and emerging narratives for journalism
Gather competitive intelligence from podcast conversations
Perform open-source intelligence (OSINT) using audio data

Frequently Asked Questions

What podcasts are included?

Currently the most popular podcasts based on the number of iTunes ratings are included (not the rating itself). The long term aim is to include as many podcasts as possible. If you think a podcast should be included please get in touch using the chat box (priority will be given to requests from subscribers).

Where does the transcription model make errors?

Generally the AI is very good at transcribing, but you might find small errors with people's names, fast speech and when people are talking over music. Long periods of silence and/or music can also introduce errors in the timestamps which can cumulate over long recordings. If you see a particularly bad transcript, let us know and we can try a different model. Some podcast providers like to change their adverts too. So if the advert was different to the one at the time of transcription, you may notice a difference. If the advert was different in length, then there might be an error of a few seconds in the timestamps.

What file format are the transcripts?

TXT, VTT and SRT with only the latter two containing timestamps. These are all ASCII files and can be opened with a text editor, e.g. Notepad.

What about paywalled/exclusive podcasts (e.g. Spotify)?

These are not yet included.

How often is Tapesearch updated?

Our database starts a refresh cycle at midnight UTC every day. Transcripts are usually ready shortly after.