Particle, the AI newsreader startup founded by former Twitter engineers, has officially pivoted to tackle a massive blind spot in the digital landscape: audio data. By launching Radar, a sophisticated podcast search engine, the company is transforming spoken conversations into structured, searchable, and highly actionable intelligence for AI agents and enterprise users.

Bridging the "Audio Blind Spot" for AI Agents

As the industry moves toward autonomous AI agents, a significant data gap has emerged. While most web-crawling agents are highly proficient at processing text, they are effectively "blind" to the vast amount of information contained within audio formats. Particle CEO Sara Beykpour notes that unless audio is transcribed and indexed, the intelligence within it remains inaccessible to programmatic tools.

Radar solves this by providing a robust API and Model Context Protocol (MCP) integration, allowing AI agents to "see" into the audio world. This enables developers to build agents that can reason about spoken content, making the spoken word just as navigable as a Wikipedia article or a news site.

Massive Scale and Granular Metadata

Radar is not just a simple transcription service; it is a massive intelligence engine. The platform currently indexes more than 130,000 podcasts, making it the largest transcribed podcast service in existence. This index includes the entirety of the Apple Top 200 podcasts across 135 different verticals, with a staggering 20,000 new episodes added to the database every single day.

The technical depth of the indexing goes far beyond raw text. Radar provides:

  • Speaker Labels: Distinguishing between guests and hosts for context.
  • Entity Recognition: Identifying specific people, companies, brands, products, and topics.
  • Rich Metadata: Tracking listener ratings, reviews, and even specific advertisement placements.
  • Self-Contained Clips: Extracting timestamped, relevant audio snippets so users can listen to specific insights without playing an entire hour-long episode.

High-Value Use Cases: From Hedge Funds to Ad Tracking

The commercial viability of Radar is already being proven by high-stakes users. Hedge funds have emerged as some of the highest-volume customers, utilizing the API to feed podcast insights into their proprietary trading agents—capturing alpha from spoken trends that text-based crawlers miss.

Beyond finance, the platform offers specialized tools for brand and media monitoring:

  • Custom Alerts: Users can configure automated alerts via Slack, email, or webhooks when specific guests discuss particular topics.
  • Ad Intelligence: A dedicated podcast ads search engine allows companies to track where they are being mentioned and how sponsorship trends evolve over time.
  • Analytical Depth: The tool provides political bias analysis, audience size estimates, and brand suitability metrics.

As Particle looks to the future, the roadmap includes expanding this audio intelligence beyond podcasts to include YouTube videos and news clips, aiming to become the definitive intelligence layer for all spoken media.

Key Takeaways

  • Solving the Audio Gap: Radar provides the necessary API and MCP layers to allow text-centric AI agents to process and understand spoken audio data.
  • Unmatched Scale: With over 130,000 podcasts indexed and 20,000 new episodes added daily, it represents one of the largest audio intelligence datasets available.
  • Enterprise-Grade Intelligence: The platform serves high-value sectors like hedge funds and advertisers by offering granular entity tracking, ad monitoring, and automated alerts.