Particle pivots to Radar, a podcast search engine built for AI agents

Particle is shifting its focus from its news-reading app to what it sees as a more lucrative opportunity: indexing the spoken conversations inside podcasts and making them discoverable. Radar, introduced Wednesday, doesn’t just transcribe podcast audio — it extracts key quotes and highlights by understanding what’s being said.

Chief executive Sara Beykpour told TechCrunch that hedge funds have become Radar’s highest-volume customers integrating directly with its API, since they want data their own agents can’t otherwise see. Journalists and researchers could use the tools too, she said, but other top-paying customers so far include AI search platforms and data resellers; Exa, a search API provider for AI agents, is among Radar’s partners.

The idea grew out of one of the news-reading app’s most popular features, which used the same underlying technology to source podcast clips for the app’s news feed. Particle’s team realized the capability was valuable but trapped inside the news reader, and as interest in AI agents grew, decided to spin it into a standalone API for podcast intelligence.

“Agents are generally blind to audio; they can’t see it unless something or someone has transcribed it,” Beykpour said, describing Radar’s ambition to eventually cover all audio, not just podcasts.

Radar’s transcriptions include speaker labels and metadata identifying the people, companies, brands, products and topics discussed, covering all of Apple’s Top 200 podcasts across 135 verticals, with 20,000 new episodes added to the index daily. The service can track mentions of specific entities across shows and send alerts — by email, Slack or webhook — when a mention occurs or as a daily or weekly digest, with filters for particular guests or topics.

Radar also surfaces pre-selected, self-contained clips with timestamps, tracks episode ads through a dedicated ads-search tool, and offers additional data products such as political-bias analysis, chart rankings, audience-size estimates, sponsorship data and brand-suitability scoring. While all of this is available through a web interface, the company positions the API and an MCP integration — letting AI agents and other businesses query the intelligence programmatically — as the core product.

Radar is priced at $29 a month per seat, with a $399-a-month plan covering 20 seats for businesses; API access is custom-priced. Particle plans to eventually extend the service beyond podcasts to other audio, including YouTube videos and news clips.

Why it matters

Radar is a bet that AI agents need a dedicated layer for audio the way they already have one for text — a gap that, if it holds, could turn podcast transcription into infrastructure rather than a niche tool.