Particle pivots to Radar: AI-powered podcast search engine built for intelligent agents

Particle, an AI newsreader platform founded by ex-Twitter staff, is charting a new direction. The company announced Radar, a podcast discovery platform that transcribes audio, extracts meaning from it, and surfaces key moments—all designed to feed AI agents with data they can't normally reach.

The business model is already working. Particle co-founder and CEO Sara Beykpour revealed that hedge funds have emerged as Radar's biggest commercial users, integrating the system via API to source intelligence for their autonomous tools. Beyond finance, AI search platforms, data resellers, and research teams have also signed on.

From news app feature to standalone product#

Radar grew out of a popular feature inside Particle's own news-reading app—one that curated podcast clips to pair with related stories. When the team recognized the untapped potential and saw momentum building around AI agents, they decided to extract and expand that capability into a standalone service.

"Agents are generally blind to audio; they can't see it unless something or someone has transcribed it," Beykpour explained. By indexing spoken conversations, Radar fills a gap that text-focused web crawlers leave open.

Scale and capabilities#

The platform currently indexes over 130,000 podcasts—claimed to be the largest transcribed podcast database available. This includes all top 200 Apple podcasts across 135 categories, with roughly 20,000 new episodes cataloged daily. Each transcript includes speaker identification and rich metadata, with the system recognizing mentions of people, companies, brands, products, and topics.

Key features include:

• Smart alerts triggered when tracked entities are discussed, delivered via email, Slack, or webhook with customizable filters
• Timestamped clip extraction of noteworthy moments listeners can watch or read
• Ads tracking that finds episodes where specific brands advertise and monitors trends
• Metadata analytics covering audience ratings, review data, episode sponsors, and estimated listener size
• Political bias analysis and brand suitability scoring

Pricing and expansion plans#

Consumer access starts at $29 monthly per user, with a $399/month team tier supporting 20 seats. Businesses and AI services negotiate custom API pricing. Beyond the web interface, the real revenue driver is the API and model context protocol (MCP) access, which lets agents and enterprises query the podcast intelligence layer programmatically.