Particle, the AI newsreader startup founded by former Twitter engineers, is shifting its focus toward podcast intelligence with the launch of Radar, a search engine designed to make spoken conversations inside podcasts searchable.
Radar does more than transcribe audio. It analyzes what speakers are discussing and can identify important quotes, highlights, people, companies, brands, products, and topics.
Hedge Funds Are Already Using Radar
Particle co-founder and CEO Sara Beykpour said hedge funds have become the highest-volume customers directly integrating Radar through its API.
Journalists and researchers can also use the service, while other paying customers include AI search platforms and data resellers. AI search API provider Exa is among Radar’s partners.
The technology originated inside Particle’s newsreader app, which used its API to find interesting podcast clips and place them alongside related news stories.
Particle eventually realized the technology could work as a standalone product. As AI agents became more widely used, the company shifted its focus toward an API that gives these systems access to information contained in audio.
Beykpour said most AI agents and services crawl text on the web and are effectively blind to audio unless someone first transcribes it.
More Than 130,000 Podcasts Indexed
Radar currently transcribes more than 130,000 podcasts, which Particle says makes it the largest transcribed podcast service in existence.
Its database includes all Apple Top 200 podcasts across 135 categories, while around 20,000 episodes are added every day.
Transcripts include speaker labels and detailed metadata. Radar also understands which people, companies, brands, products and topics are being discussed.
Users can track these entities and receive alerts whenever they are mentioned. Notifications can arrive immediately or as daily or weekly digests through email, Slack or webhooks.
Users can also apply filters, such as receiving an alert only when a particular guest discusses a specific topic or limiting searches to top podcasts.
Radar Can Search Quotes, Clips and Ads
Radar can extract relevant standalone clips with timestamps, allowing users to either listen to the original audio or read what was said.
Particle also preselects notable clips for users who do not want to listen to an entire podcast or read a full summary.
The platform can track topics, mentions, listener ratings and reviews, episode advertisements and other podcast data.
Radar also includes a dedicated podcast advertising search engine that can find every episode where a company advertises and track those appearances over time.
Particle also sees potential in political bias analysis, chart ranking data, audience size estimates, sponsorship information and brand suitability tools.
Pricing and Future Expansion
Radar is available through a web interface, but Particle says its main product is the API and MCP, which allow AI agents and businesses to access the same podcast intelligence programmatically.
Radar costs $29 per month per seat, while its business plan costs $399 per month for 20 seats. API pricing is customized according to customer requirements.
Particle plans to eventually expand Radar beyond podcasts to other forms of audio and video, including YouTube videos and news clips.
The post This AI Search Engine Can Find Anything Said in Podcasts appeared first on ProPakistani.