Radar Converts Podcasts into Searchable Content for AI Agents

Particle, the startup founded by former Twitter engineers, is shifting gears to pursue a potentially lucrative opportunity: indexing spoken conversations within podcasts and improving their accessibility. On Wednesday, the firm unveiled Radar, a podcast search engine that not only transcribes audio but also understands its context, enabling it to highlight important quotes and key moments.

This innovation opens up business prospects, having already piqued the interest of hedge funds looking for otherwise inaccessible data, as highlighted by Particle co-founder and CEO Sara Beykpour.

“Hedge funds represent our largest clients that are directly integrating with our API,” Beykpour explained to TechCrunch. While the tools are available for journalists and researchers, high-paying clients also include AI search platforms and data resellers like Exa, a search API provider for AI agents that partners with Radar.

The idea came from one of the most valued features of the Particle news-reading app, which used its API to collect captivating podcast snippets to accompany related news articles in its feed.

The Particle team saw the product’s promise but realized it was somewhat limited within the confines of the news reader. As interest in AI agents surged, the company decided to redirect its focus towards creating an API specifically for its podcast intelligence services.

“We aim to cover all new media intelligence and audio intelligence through this API. The appeal of this field lies in the fact that most API agents and services concentrate on text and web crawling. We are introducing that layer with audio,” Beykpour stated. “Agents generally overlook audio; they can’t access it unless it’s transcribed by someone or something.”

With Radar, the company transcribes more than 130,000 podcasts, making it the largest podcast transcription service available. This includes all podcasts from the Apple Top 200 across 135 categories, adding 20,000 episodes to Radar’s index daily.

The transcriptions feature speaker labels and rich metadata, as Radar identifies the discussed entities—like people, companies, brands, products, and topics.

Additionally, it can track mentions of these entities across various podcasts and send alerts whenever they occur, whether in real-time or as daily or weekly summaries.

The alerts, available via email, Slack, or webhook, can be customized with specific filters. Users can set Radar to notify them only when certain guests appear and discuss selected topics. The search functionality can also be fine-tuned to focus solely on top podcasts.

Radar is adept at extracting valuable, self-contained clips complete with timestamps, enabling users to both listen to and read the insights shared.

“We have pre-selected noteworthy clips, so if you can’t listen to the entire podcast or prefer not to read a summary, this provides an excellent way to understand the gist of the podcast,” Beykpour noted.

Radar can also monitor the subjects discussed in each podcast, identify the people or topics being addressed, and track listener ratings and reviews, including ads featured in the episodes. It also includes a dedicated search engine for podcast advertisements that can pinpoint every episode where a particular company promotes its products and analyze trends over time. This capability introduces extra monetization opportunities alongside tools that provide political bias evaluations, chart rankings, audience size estimates, sponsorship details, and brand suitability insights.

All of these features are available through Radar’s web interface, but the primary offering is the API and MCP, which enables AI agents and other businesses to programmatically access the same intelligence.

Radar is priced at $29 monthly per seat, with a $399 monthly plan for businesses that includes 20 seats. API users will receive custom pricing tailored to their specific needs.

In the future, Radar plans to broaden its services beyond podcasts to include other audio formats like YouTube videos and news segments.