Particle launched Radar as a standalone API and MCP product for podcast search.
Why it matters: It gives developers and AI agents a way to query podcast audio as structured data, not just listen to it. That could matter for media companies, ad-tech teams and data buyers looking for searchable audio catalogs and monitoring tools.
- Particle says Radar indexes more than 130,000 podcasts and adds about 20,000 episodes a day.
- The product exposes speaker-labeled transcripts, quotes, highlights, sponsorship intelligence and advertising data through REST, MCP and a real-time firehose.
- TechCrunch reported the launch and said Particle sees hedge funds, AI search platforms and data resellers as customers.
- Particle’s pricing page says developers can access the service through REST, MCP and a real-time firehose.
Particle has turned the podcast tooling inside its news app into Radar, a standalone data product for AI agents and developers. The company says the service transcribes audio, extracts quotes and highlights, and makes that material searchable through an API and model context protocol, or MCP, which lets AI systems connect to external tools. TechCrunch reported the launch.
Particle’s pricing page says Radar includes semantic search, full-text search, rich metadata and access through REST, MCP and a real-time firehose. In plain terms, that means customers can query podcasts directly, pull results into software and stream updates as new episodes arrive.
The company says Radar covers more than 130,000 podcasts and adds roughly 20,000 episodes a day. Those are Particle’s own figures. Its pitch is that audio is still hard for machines to use unless it is transcribed and tagged, so Radar is meant to make spoken content searchable for humans and agents alike.
Particle CEO Sara Beykpour said in the launch materials that agents are often blind to audio unless it has been transcribed. TechCrunch also reported that Particle has seen hedge funds as its highest-volume direct API customers, alongside interest from AI search platforms and data resellers. Particle’s LinkedIn post said the podcast intelligence API is built for humans and agents and can surface mentions, semantic queries, clips, ad copy and metadata. The company’s blog says alerts support through the API and MCP is live.
Particle first introduced podcast clips in its news app in February, then packaged the same system as an API product. The new launch separates that capability from the consumer app and sells it as infrastructure for search and data workflows.
By the numbers
- 130,000+ - podcasts Particle says Radar indexes
- 20,000 - episodes Particle says it adds per day
Yes, but: Most of the launch details and usage claims still come from Particle, so traction outside the company is not independently established in the available reporting.
What's next: Particle’s blog says alerts support through the API and MCP is live, suggesting the company will keep expanding the product.