Particle, the startup founded by former Twitter engineers, has launched Radar, a podcast search engine designed to make spoken conversations discoverable and usable by AI agents. The product transcribes audio, identifies entities and extracts clips and highlights so applications can query podcast content programmatically.
How Radar works
According to Particle, Radar transcribes more than 130,000 podcasts, including all the Apple Top 200 podcasts across 135 verticals, and adds about 20,000 episodes to its index every day. Transcriptions include speaker labels and metadata that identify people, companies, brands, products and topics discussed in each episode.
Radar can pull out self-contained clips with timestamps and present both audio and text for those excerpts. The system also tracks mentions of specific entities across podcasts and can send alerts when a target appears. Alerts can be delivered via email, Slack or webhook, and users can filter them by guest, topic or by limiting results to top podcasts.
Business use and integrations
Particle says hedge funds have been among the highest-volume customers integrating Radar’s API, a use case the company highlighted through comments by co-founder and CEO Sara Beykpour. Other paying customers include AI search platforms and data resellers; Exa is listed as a partner that uses Radar’s search API for agents.
Beyond search and alerting, Particle describes additional product features such as a podcast ads search engine, political bias analysis, chart rankings data, audience size estimates, sponsorship tracking and brand suitability tools. The company positions its API and MCP as the primary products for programmatic access.
Radar is offered through a web interface and an API. Pricing starts at $29 per month per seat, with a $399-per-month business plan that includes 20 seats; API customers receive custom pricing. Particle says it plans to expand Radar’s audio indexing beyond podcasts to formats like YouTube videos and news clips.
Original source: TechCrunch AI