Highlights

Particle — the AI newsreader startup founded by former Twitter engineers — has launched Radar, a podcast search engine that transcribes and semantically analyzes spoken content. Radar indexes over 130,000 podcasts, including the Apple Top 200 across 135 verticals, adding roughly 20,000 episodes daily. It produces speaker-labeled transcripts, identifies entities (people, companies, brands, topics), and can surface notable clips with timestamps. This capability makes audio accessible to AI agents that otherwise can’t interpret raw audio, unlocking new business uses for hedge funds, AI search platforms, and data resellers.

Sentiment Analysis

• Overall sentiment: positive — Radar presents a practical solution to a clear limitation in current agent and search ecosystems by converting audio into structured, searchable intelligence. The product’s strengths include breadth (130,000+ podcasts indexed), depth (speaker labels, entity recognition, metadata), and delivery options (web UI, alerts, and a programmatic API). These features make it attractive to customers who need reliable audio-derived data, such as hedge funds, AI platforms, and researchers. The market response is favorable, reflected by early integrations and business interest.
  
  
  

  75%

  
  
  
  

Article Text

Particle, a startup created by former Twitter engineers and previously known for an AI-powered newsreader app, has pivoted to a focused product: making podcast audio discoverable and useful to AI-driven systems. Their new offering, Radar, does more than transcribe audio — it interprets content, extracts key quotes and highlights, and packages the results so that both humans and automated agents can find and act on spoken information.

The origin of Radar traces back to a popular feature in Particle’s earlier app: the ability to surface compelling podcast clips alongside related news stories. That capability demonstrated the value of structured podcast data, but it was constrained within the original newsreader. As interest in autonomous agents and programmatic data access grew, Particle repositioned the technology as a standalone API product focused on audio intelligence.

Radar’s corpus is substantial: the service transcribes more than 130,000 podcasts, including all shows in Apple’s Top 200 across 135 vertical categories, and adds roughly 20,000 new episodes to its index each day. Transcriptions include speaker labels and rich metadata. Radar recognizes entities — such as individuals, companies, brands, products, and topics — and can follow mentions across episodes. Users can receive alerts when specific entities or subjects are mentioned, either in real time or as daily or weekly digests, delivered via email, Slack, or webhooks. These alerts are filterable, allowing users to target particular guests, topics, or subsets like top-ranked podcasts.

Beyond search and alerts, Radar can extract concise, self-contained clips with timestamps so users can quickly listen to or read the exact segment where a topic is discussed. This clip extraction is useful for people who need quick access to salient moments without consuming entire episodes. The platform additionally tracks listener ratings and reviews, episode ad content, and topic trends. One feature targets advertising specifically: a podcast ads search engine that locates every instance where a company runs an ad and shows how that advertising presence evolves over time.

These capabilities create multiple commercial opportunities. Early customers with strong demand include hedge funds that want structured audio data for their agent-driven research workflows, AI search providers, and data resellers. Particle’s CEO notes hedge funds are among the highest-volume direct API customers. Other potential value-adds — and revenue streams — include political-bias assessments, chart-ranking metrics, audience-size estimates, sponsorship and brand-suitability analysis, and dedicated ad-tracking services.

While Radar offers a web interface for browsing transcripts, clips, and alerts, its primary product is the programmatic access layer: an API and management console that let AI agents and enterprise systems incorporate audio intelligence into automated processes. This approach addresses a common limitation: many crawlers and agent platforms focus on text and are effectively blind to raw audio unless it has been transcribed and structured. Radar fills that gap by supplying the audio layer of intelligence to agents and services.

Pricing for Radar’s web product is tiered: $29 per month per seat for individual users and a $399-per-month business tier that includes 20 seats. API customers receive custom pricing based on volume and integration needs. Looking ahead, Particle plans to broaden Radar’s scope beyond podcasts to include other audio sources such as YouTube videos and news clips, extending the same transcription, entity extraction, and metadata features to a wider set of audio content.

Key Insights Table