Particle Launches Radar to Make Podcast Conversations Searchable for AI Agents

Particle, an AI newsreader startup established by former Twitter engineers, has unveiled Radar, a specialised podcast search engine and intelligence platform. The tool transcribes spoken audio and extracts contextual meaning, making podcast discussions searchable and usable for software applications and artificial intelligence agents.

What Happened

Particle shifted focus from its news-reading application to address the lack of searchable audio content on the internet. On Wednesday, the startup introduced Radar to index spoken audio across more than 130,000 podcasts, covering the Apple Top 200 shows across 135 verticals. According to the company, roughly 20,000 new episodes are added to the platform’s index each day.

Rather than merely converting speech to text, Radar identifies discussed entities such as individuals, companies, products, brands, and topics, attaching speaker labels and metadata to the transcripts. It also provides self-contained audio clips with timestamps, allowing users to listen to or read specific segments without going through an entire episode.

Key Highlights

• Targeting the Audio Blind Spot: Particle co-founder and CEO Sara Beykpour noted that most web-crawling services focus primarily on text, leaving AI agents largely blind to audio data unless it is transcribed.
• Broad Enterprise Adoption: Top-paying users include hedge funds looking for alternative data, data resellers, and AI search platforms such as Exa, an API provider for AI agents.
• Alerts and Monitoring: Users can configure custom alerts through Slack, email, or webhooks based on specific topics, appearances by selected guests, or mentions filtered to top-ranking podcasts.
• Ad and Intelligence Tracking: Radar includes an ad search engine that monitors corporate sponsorships across episodes over time, alongside analytics for brand suitability, political bias, listener ratings, audience estimates, and chart performance.
• Access and Pricing: In addition to an API and MCP access for automated queries, Radar offers a web interface priced at $29 per seat monthly, a business tier at $399 per month for 20 seats, and custom tiers for API usage.

Why This Matters

Automated web systems and generative AI tools rely heavily on written web pages to gather information. Because spoken audio remains largely locked behind audio files, platforms like Radar supply a structured text and metadata layer that allows automated agents and financial analysts to access spoken discussions, corporate commentary, and commercial trends that were previously unindexed.

What to Watch Next

Particle plans to broaden the scope of Radar beyond traditional podcasts. Future updates are intended to process and index other forms of spoken media, including YouTube videos and broadcast news clips.

Frequently Asked Questions

What is Particle’s Radar?

Radar is a podcast search and intelligence tool created by Particle that transcribes spoken audio, labels speakers, and indexes mentions of people, brands, companies, and topics.

Who is currently using the platform?

According to Particle, hedge funds represent the highest-volume customers integrating via the API. Other users include AI search platforms, data resellers, and researchers.

How much does Radar cost?

Radar is priced at $29 per month per seat for individual web access and $399 per month for a 20-seat business plan. Custom pricing models are used for direct API customers.

Source: TechCrunch reporting on Particle announcements.