Radar makes podcasts searchable — and usable by AI agents
TechCrunch Sarah Perez
Particle launched Radar, a podcast search tool that can transcribe and understand audio. It turns podcasts into data AI agents can actually use, and hedge funds already care.
Based on reporting by TechCrunch, Sarah Perez — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Particle, the AI newsreader startup from former Twitter engineers, is moving into something a lot more commercially pointed than a clever feed feature: podcast intelligence. On Wednesday, it launched Radar, a search engine for podcasts that doesn’t just turn speech into text. It tries to understand the content, pull out meaningful quotes, and make the whole mess of spoken conversation searchable.
That matters because Particle says the audience is already there, and it isn’t only journalists or researchers. Sara Beykpour, the company’s co-founder and CEO, said hedge funds have been the biggest direct API customers so far. AI search companies and data resellers are also paying attention, and Exa, a search API provider for AI agents, is listed as a partner.
The product came out of a feature inside Particle’s news app, where podcast clips were already being surfaced next to related stories. That helped the team see the demand, but also the limits of keeping the idea inside a news reader. So Particle shifted toward the thing that could stand on its own: an API built for agents and other businesses that need audio data instead of web text.
Radar now transcribes more than 130,000 podcasts, including all the Apple Top 200 podcasts across 135 verticals, and adds 20,000 episodes a day. The system labels speakers, tracks people, companies, brands, products and topics, and can send alerts by email, Slack or webhook when specific names or themes appear. Users can filter those alerts, narrow searches to top podcasts, or pull out self-contained clips with timestamps when they just want the useful bit without listening to the whole episode.
There’s also a dedicated search engine for podcast ads, so companies can find every episode where a brand advertises and watch how that changes over time. Particle says the web interface is only part of the story. The real product is the API and MCP, which let AI agents and other systems tap into the audio intelligence directly. Pricing starts at $29 a month per seat, with a $399 monthly business plan that includes 20 seats, while API pricing is custom. The company plans to expand beyond podcasts eventually, with YouTube videos and news clips next on the list.
My take — AI-written commentary, not fact-checked reporting
This is the sensible kind of AI play: stop pretending audio is “content” and start treating it like a database. The bigger signal is that the buyers with money are the ones who want machine-readable media, not yet another app that summarizes a podcast nobody finished anyway. Also, hedge funds leading the line is a nice reminder that “knowledge work” usually means “who can pay for the transcript first.”
Read more about this at: TechCrunch
