Particle launches Radar, making podcasts searchable for AI agents

Particle, an AI newsreader startup founded by former Twitter engineers, has pivoted to focus on indexing and making spoken conversations in podcasts discoverable and usable by AI agents. The company launched Radar, a podcast search engine that transcribes audio, understands its meaning, and extracts key quotes and highlights.

Radar's ability to make audio content accessible to AI agents, which are typically focused on text, has attracted significant interest from hedge funds, AI search platforms, and data resellers. Particle CEO Sara Beykpour explained that hedge funds are the highest-volume customers directly integrating with Radar's API, highlighting the business potential of providing audio intelligence.

The product evolved from a popular feature in Particle's news-reading app that sourced podcast clips. Recognizing the value and the growing movement around AI agents, the company decided to focus on building an API for its podcast intelligence. Radar currently transcribes over 130,000 podcasts, adding 20,000 episodes daily, and includes speaker labels and rich metadata.

The service allows users to track mentions of entities like people, companies, and brands, and receive custom alerts via email, Slack, or webhook. It can also extract self-contained audio clips with timestamps, providing a way to quickly grasp podcast content without listening to or reading a full summary. Additional features include a dedicated podcast ads search engine, political bias analysis, and audience size estimates.

While Radar offers a web interface, its primary product is the API and MCP, enabling programmatic access for AI agents and businesses. Pricing starts at $29 per month per seat, with a $399 monthly plan for businesses. Particle plans to expand Radar's capabilities to other audio formats like YouTube videos and news clips in the future.