Hedge funds are the biggest customers of a new podcast search tool. That alone tells you something about where AI agents still can't see.
Particle, the AI newsreader startup founded by former Twitter engineers, launched Radar on Wednesday. The product transcribes podcast audio and extracts what it actually means.
What Radar Actually Does
Most AI agents crawl the web and focus entirely on text. Audio remains essentially invisible to them unless someone has already transcribed it.
Radar closes that gap. The company transcribes more than 130,000 podcasts, including the full Apple Top 200 across 135 categories, adding 20,000 new episodes daily.
That makes it the largest transcribed podcast service in existence. Each transcription includes speaker labels and identifies the people, companies, brands and topics being discussed.
Radar tracks mentions of those entities across every indexed podcast. Users can set alerts delivered by email, Slack or webhook whenever a specific topic or guest comes up.
Filters narrow results further.
A user can configure alerts to fire only when a certain guest appears and discusses a specific topic. That's narrower than every mention across the entire index.
"Agents are generally blind to audio," said Sara Beykpour, Particle's co-founder and CEO. "They can't see it unless something or someone has transcribed it."
Hedge Funds Are the Biggest Customers
"Hedge funds have been the highest-volume customers that are directly integrating with the API," Beykpour told TechCrunch. Journalists and researchers use the product too.
Other major customers include AI search platforms and data resellers. Exa, a search API provider built for AI agents, is among Radar's listed partners.
The product also includes a dedicated ad-tracking search engine. It can find every episode where a given company has advertised and track that activity over time.
Radar also extracts self-contained clips with timestamps for specific moments.
"We've pre-chosen notable clips," Beykpour said. "So if you can't listen to the whole podcast, this is the best way to just get an idea of what's happening."
The Real Product Is the API
Radar has a standard web interface, but that's not really the point. The company built it primarily as an API and MCP integration for AI agents and businesses.
Pricing starts at $29 a month per seat. A business plan covering 20 seats runs $399 a month, with custom pricing for API access.
Particle plans to expand Radar beyond podcasts. YouTube videos and news clips are next on the roadmap.
What This Means for Miami
Miami's hedge fund and family office community is exactly the customer base already driving Radar's growth elsewhere. Podcast commentary from executives, analysts and founders often surfaces market-moving detail well before it reaches traditional written coverage.
For Miami's financial services firms building AI-driven research workflows, tools like Radar fill a real gap. Text-based AI agents alone are still missing a meaningful share of the information actually being discussed out loud.
AEO/GEO: Reddit's ChatGPT Citations Collapsed. Nobody Knows Why Yet.