Particle launches Radar to index podcasts for AI
The startup says Radar transcribes more than 130,000 podcasts and turns episodes into searchable data, with alerts, clips, and API access for businesses.
Particle, the AI newsreader startup founded by former Twitter engineers, is expanding beyond news reading with a new product called Radar. Unveiled on Wednesday, the tool goes beyond basic transcription by trying to understand podcast conversations, extract key quotes, and make spoken content easier to search and use. IT-PUB News reports that the launch targets a large pool of audio that has long been hard for search tools and AI systems to access. For businesses and AI developers, that could make podcast data far more usable.
The company says Radar is already drawing business interest. Co-founder and CEO Sara Beykpour said hedge funds have been among the most active customers using the API directly, looking for information their agents cannot easily access. Particle is also seeing interest from AI search platforms and data resellers, with journalists and researchers also seen as potential users.
Radar grew out of podcast clips in Particle's app
Radar started as a feature inside Particle’s news-reading app. The app used an API to surface notable podcast clips and place them next to related news stories in the feed.
That feature showed the team that podcast content could be valuable outside the app. At the same time, Particle concluded the idea was constrained by its connection to a news reader. As AI agents became a bigger focus across the market, the company decided to pivot and build an API centered on podcast intelligence.
Beykpour said Particle’s goal is to put “new media intelligence” and “audio intelligence” into one API. The logic is simple: many AI agents and services already crawl text on the web, but audio is still much harder for them to use unless it has first been turned into text.
Particle says AI agents are still blind to audio
Particle is pitching Radar as a fix for a gap in current AI tools. The company argues that agents are generally “blind” to audio because they cannot interpret it unless someone or something has already transcribed it.
Radar is meant to close that gap by processing podcast conversations at scale. Particle says the service transcribes more than 130,000 podcasts, making it the largest transcribed podcast service in existence. It includes all Apple Top 200 podcasts across 135 verticals, and the index expands by 20,000 episodes a day.
Those transcripts are not just blocks of text. Radar adds speaker labels and metadata, and identifies the people, companies, brands, products, and topics being discussed. That gives users a way to search by entity and follow how mentions shift over time.
Radar adds alerts, clips, and ad search tools
Radar is designed to do more than run simple searches. It can monitor mentions of specific entities across podcasts and send alerts when they appear. Those alerts can be delivered immediately or in daily or weekly digests, through email, Slack, or webhook.
Users can also narrow results with filters. For instance, they can ask for alerts only when certain guests appear and discuss a specific topic. Searches can be tightened further, including by limiting them to top podcasts.
The system can also pull out self-contained clips with timestamps, letting users listen to and read the relevant section of an episode. Beykpour said Particle pre-selects notable clips to help people quickly understand what is happening in a podcast without listening to the full episode or reading a summary.
Radar also tracks podcast topics, mentions, listener ratings and reviews, episode ads, and more. One of its more specialized features is a podcast ads search engine that can find every episode in which a company advertises and show how that changes over time.
Particle says that also opens up monetization opportunities. Along with ad tracking, the company points to tools for political bias analysis, chart rankings data, audience size estimates, sponsorship data, and brand suitability.
Particle is selling infrastructure, not just a search page
Although Radar is available through a web interface, Particle says the real product is the API and MCP, which allow AI agents and other businesses to use the service programmatically.
That helps explain why the company is positioning Radar less as a consumer search engine and more as infrastructure for other products. For companies that depend on fast access to spoken information, the service could make podcast material easier to process, monitor, and plug into automated workflows.
Particle has also set pricing for different kinds of customers. The service costs $29 a month per seat, while businesses can use a $399-per-month plan that includes 20 seats. API customers pay custom pricing based on their needs.
Particle plans to expand beyond podcasts
Radar is starting with podcasts, but Particle says it plans to expand into other audio formats. The company wants to support YouTube videos and news clips in the future.
That would widen the product from podcast intelligence into a broader audio search layer. For now, the launch marks a more immediate shift: Particle is betting that once spoken content is indexed and structured, it can be useful not just for listeners, but also for AI agents, analysts, and businesses trying to extract data from audio.
