Radar makes podcasts searchable — and usable by AI agents

2 weeks ago 32

Particle, the AI newsreader startup founded by erstwhile Twitter engineers, is shifting its absorption to a perchance much lucrative idea: indexing the spoken conversations buried in podcasts and making them discoverable. On Wednesday, the institution introduced Radar, a podcast hunt motor that not lone transcribes podcast audio but besides understands what it means, enabling it to propulsion retired cardinal quotes and highlights.

The solution has concern potential, arsenic it’s already attracted involvement from hedge funds looking for information that their agents can’t see, explains Particle co-founder and CEO Sara Beykpour.

“Hedge funds person been the highest-volume customers that are straight integrating with the API,” Beykpour told TechCrunch. While journalists and researchers could besides marque usage of the tools, different top-paying customers person included AI hunt platforms and information resellers. (The hunt API supplier for AI agents, Exa, for instance, is among Radar’s partners.)

The thought itself stemmed from 1 of the Particle news-reading app’s astir beloved features. The app had utilized an API to source absorbing podcast clips that it past included alongside related quality stories successful the app’s feed.

Particle’s squad realized the product’s value, but besides that it was somewhat trapped successful the quality reader. As the question astir AI agents began to summation steam, the institution decided to pivot and absorption connected gathering an API for its podcast quality product.

“Our imaginativeness is truly to person each caller media quality and each audio quality successful that API. One of the reasons wherefore it’s an absorbing abstraction is that astir API agents and services crawl the web and they’re focused connected text. We are providing that furniture with audio,” Beykpour said. “Agents are mostly unsighted to audio; they can’t spot it unless thing oregon idiosyncratic has transcribed it.”

With Radar, the institution transcribes much than 130,000 podcasts, making it the largest transcribed podcast work successful existence. This includes each the Apple Top 200 podcasts crossed its 135 verticals, with 20,000 episodes added to Radar’s scale daily.

The podcast transcriptions see talker labels and affluent metadata, arsenic Radar understands the entities — people, companies, brands, products, and topics — being discussed.

It’s besides capable to way mentions of these entities crossed podcasts and nonstop alerts whenever they travel up, either erstwhile the notation occurs oregon arsenic a regular oregon play digest.

The alerts, which tin beryllium delivered via email, Slack, oregon webhook, tin beryllium customized with filters. These fto users configure Radar to lone nonstop alerts erstwhile definite guests look and sermon a peculiar topic, for instance. The hunt tin besides beryllium narrowed successful different ways specified arsenic limiting it to apical podcasts only.

Radar tin extract applicable self-contained clips, with timestamps, allowing users to some perceive to and work the comments made.

“We’ve pre-chosen notable clips, truthful if you can’t perceive to the full podcast and you don’t privation to work a summary, this is the champion mode to conscionable get an thought of what’s happening successful that podcast,” Beykpour noted.

Radar tin besides way the topics mentioned successful the podcast, who oregon what was mentioned and when, listener ratings and reviews, the episode’s ads, and more. There’s adjacent a dedicated podcast ads hunt motor that tin find each occurrence wherever a fixed institution advertises and way however it trends implicit time. This diagnostic has further monetization potential, alongside different tools offering governmental bias analysis, illustration rankings data, assemblage size estimates, sponsorship data, and marque suitability.

While each of this is disposable done Radar’s web interface, its existent merchandise is the API and MCP, which allows AI agents and different businesses to pat into this aforesaid quality programmatically.

Radar is priced astatine $29 a period per seat, with a $399-per-month program for businesses that includes 20 seats. API users person customized pricing, based connected their needs.

In the future, Radar plans to grow the work beyond podcasts to enactment different forms of audio, specified arsenic YouTube videos and quality clips.

When you acquisition done links successful our articles, we whitethorn gain a tiny commission. This doesn’t impact our editorial independence.

Read Entire Article