For agentsread agents.md

Podcast intelligence
for agents.

150,000+ podcasts, fully transcribed, diarized, and speaker-identified, within minutes after airing.

StatusLoading…Total podcasts150,243Episodes processed (24h)72,280DiarizedSpeaker IdentifiedEntity extractedTopicsGuestsAdsTranscriptsSemantic Search

MCP Server: Connect your agent in one command.

claude mcp add --transport http particle https://mcp.particle.pro

Select Partners

Search full transcripts by entity, keyword, or semantic match.

Match a canonical entity, an exact keyword, or a semantic concept — get back speaker-labeled transcript windows, ranked by relevance, across every indexed show and episode.

$ curl -G 'https://api.particle.pro/v1/podcasts/mentions' \
    --data-urlencode 'entity_id=anthropic' \
    --data-urlencode 'limit=3' \
    -H "X-API-Key: $PARTICLE_API_KEY"
NYC NOW cover art
NYC NOWMayor Mamdani Promised to Rethink Policing. Six Months Later, the NYPD Is Growing02:27 · advertiser
“When the Pope released his AI encyclical, the co-founder of Anthropic sat beside him and gave a speech.”
Power Lunch cover art
Power LunchPower Lunch 6/8/2603:35 · Tom Lee · guest
“…Google is gonna raise eighty billion, plus SpaceX seventy-five, OpenAI maybe a hundred, you know, Anthropic seventy-five, and, and then Meta, of course, filed as well.”
The Exchange cover art
The ExchangeApple’s Big Announcements, The Bull Case for Chips, and Under-the-Radar AI Plays15:26 · Leslie Picker · correspondent
“You saw Anthropic make the move of basically saying, ‘None of these shenanigans.’”

Rich episode transcripts and metadata

Every transcript is available with full diarization and speaker labels. Rich metadata for each podcast and episode is also available — canonical entities for guests, hosts, mentioned people and companies, topics, ranks and reviews, estimated audience size, brand-safety alignment, political-bias ratings, and more.

SpeakerBrian Vitale
RoleHost
ShowRPG Site · Tetracast
Episode#445
SegmentFFVII Revelation Trailer Reactions
TopicVideo Games · Role-Playing Games
Engagement78 / 100
Brand safetySafe

Sponsorship intelligence, mapped to canonical brands.

Track every ad, every read, every placement — pre-roll vs mid-roll, host-read vs pre-recorded, offer URLs included. Updated daily across the full index.

$ curl -G 'https://api.particle.pro/v1/podcasts/advertising/leaderboard' \
    --data-urlencode 'limit=10' \
    -H "X-API-Key: $PARTICLE_API_KEY"
#BrandPodcastsEpisodesAds
1Grainger2,35719,15233,153
2Progressive2,18915,23524,516
3Mint Mobile2,48716,72718,893
4Shopify2,27715,63617,869
5Lowe's2,15010,51215,160
6BetterHelp1,67111,70013,507
7Public3835,79213,190
8FanDuel7289,74612,839
9SpinQuest7183,15112,684
10LifeLock7588,64211,061

Data for June 2026 shown.

Every Top 200 ranked show. Every vertical.

Coverage of the entire Apple Podcasts Top 200 rankings across 130+ English-language verticals, plus the long tail of independent shows worth tracking.

$ curl -G 'https://api.particle.pro/v1/podcasts/rankings' \
    --data-urlencode 'limit=200' \
    -H "X-API-Key: $PARTICLE_API_KEY"

Data for June 2026 shown.

Get notified the moment an entity is mentioned or a guest appears.

Track brands, people, and topics as canonical entities — plus guest appearances — across every podcast. Choose a cadence: as-it-happens, daily, or weekly digests.

As-it-happens

A near real-time email the moment a tracked entity or guest appears in a new episode, with clip, transcript, and metadata attached.

Daily digest

A single email each morning with tracked entity mentions or guest appearances from the prior day, ranked by audience reach and popularity.

Weekly digest

Like the daily digest, but for the week. Perfect for tracking people or companies that are named often.

  • Email
  • Slack
  • Webhooks

Built for agents.

A person at the keyboard? Connect over the Model Context Protocol — Particle drops into Claude, Cursor, and any MCP client. An agent with a wallet and no key? Pay per request with x402 or MPP. Particle also ships inside Firecrawl, Exa's research agent, the Lightfield CRM, and TinyFish — with Parallel coming soon.

Ask in plain language — your agent does the rest.

how much are hyperscalers spending on AI in 2026?

calling particle…

Notes on the Week Ahead05/04 · 05:34

"…the largest five hyperscalers now expected to spend roughly seven hundred and twenty-five billion dollars in twenty twenty-six…"

David Kelly · Host

+ 5 more across 5 shows

Get started in one command:

claude mcp add --transport http particle https://mcp.particle.pro

Or let the agent set itself up:

Read https://api.particle.pro/agents.md and connect me to Particle.
Go to MCP docs

No account, no key — your agent pays per call.

GET /v1/podcasts/search?q=hyperscaler ai capex

402 Payment Required · $0.01 USDC on Base

x402PAYMENT-REQUIRED → PAYMENT-SIGNATURE → PAYMENT-RESPONSE

MPPWWW-Authenticate: Payment → Authorization: Payment → Payment-Receipt

200 OK · settled

  • Every endpoint and MCP tool, over x402 or the Machine Payments Protocol.
  • A call that fails is never charged. Send an API key instead and the request takes the normal auth path.
  • Listed in the Coinbase x402 Bazaar.

Also available inside

Particle is a provider on Firecrawl.

Agents using Firecrawl’s Alexandria can search Particle’s podcast transcripts, episodes, and guests alongside the live web.

Go to Firecrawl

Particle connects to Exa’s research agent.

Access the power of Particle's Podcast Intelligence through Exa's research agent with Exa Connect.

Go to Exa Connect

Particle is a connector on Lightfield.

Bring Particle’s Podcast Intelligence into the Lightfield CRM: enrich your prospects with podcast data to understand everything about your customer.

Go to Lightfield

Connect your Particle account to TinyFish.

Particle is a founding partner in the TinyFish Data Partners Alliance. Link your Particle account and your TinyFish agents can work with podcast transcripts, episodes, and guests.

Go to TinyFish

Coming soon: Particle on Parallel.

Soon, agents built on Parallel will be able to reach Particle’s Podcast Intelligence alongside the web.

Go to Parallel

Radar: The Podcast Search Engine

All Particle Podcast Intelligence customers get access to Radar, a web app built on the data that the API serves. Spot trends, see everything — transcripts, entities, ads, alerts — in the browser, no need to build.

Mentions per week

“data centers” · all podcasts

↑ 20× in six months

Radar’s Trends page. Weekly mentions of “data centers” across all podcasts, from 232 in the week of Feb 23 to 4,771 in the week of Aug 17, 2026.

Search what was actually said

Search every transcript by topic, brand, or person, then jump to the moment it was said with the speaker labeled.

See where brands advertise

Pull any brand and read its podcast footprint — which shows, how many placements, host-read or pre-recorded, and the ad copy word for word.

Follow guests and trends

Track who went on which show this week, watch topics climb across the charts, and favorite the brands and people you follow.

The full stack of podcast intelligence.

Speaker-identified transcripts, semantic search, topical clips, sponsorship intelligence, and an MCP server — under one API key, or with no key at all via x402 or MPP.

ParticlePod EngineListen Notes
Coverage
Fully enriched podcasts with full transcripts + speaker labelsEvery episode——
Full transcriptsEvery episodeCapped per plan—
Speaker diarizationEvery episode——
Speaker identified and labeledEvery episode——
Word-level timestampsEvery episode——
FreshnessNew episodes available within minutesUnknownUnknown
Search
EntityCanonical entities for people, companies, topics20+ filtersPeople & topics
SemanticIncludedVector embeddings—
KeywordExact + phraseTranscriptsMetadata only
Hybrid (keyword + semantic)Combined——
AlertsEntity & guest-appearance alertsGuest appearances—
Clips & segments
Topical clips with audio + transcript windowAuto-generated—Audio only
Fully segmented chaptersEvery episode——
Embeddable HTML clips with audioHTML widget with timed text—HTML widget
Metadata
TopicsPer episodePer podcastGenres only
RatingsUser ratingsUser ratingsListen Score
RankingsIncludedIncluded—
ReviewsFull text and ratingsCounts & ratings—
Brand safety ratingsIAB GARM 12 categories——
Political bias ratingsIncludedIncluded—
Distribution & social graphDirectories, socials, video, audience sizesApple, Spotify, 7 socialsSocials only
Sponsorship
Ad detectionPer ad read——
Canonical product & company mappingResolved——
Ad leaderboardTop sponsors——
Co-occurrencesSponsor pairs——
Access
RESTYesYesYes
MCP server for AI agentsIncludedIncluded—
Pay per request with x402 or MPP (no API key)USDC on Base, every endpoint & MCP tool——
Enterprise firehose accessAvailableS3 bucketAvailable
Enterprise data snapshotsAvailableUnknownUnknown

Select your plan.

Loading plans…

Common questions.

If something isn’t here, the docs probably have it — or reach out and we’ll add it.

What is Particle?
Particle is a B2B data platform that exposes real-time company and podcast intelligence as a single API for developers, brands, and AI agents.
What APIs does Particle offer?
Particle offers two API surfaces: Company Intelligence (profiles, earnings, SEC filings, and reference data for public companies) and Podcast Intelligence (transcription, speaker diarization, and semantic search across thousands of podcasts).
Who uses Particle?
Developers building agentic and analytical products, brands and advertisers measuring earned media and podcast mentions, and research teams sourcing alternative data signals.
How do I get started?
Sign up at platform.particle.pro/register — you'll get a free $10 starter credit (about 1,000 requests). Create an API key and you're set to call the REST API directly, or wire up the MCP server for your agent in one command.
How can I access the data?
Four ways. The REST API with an API key, for scripts, backends, and productization. The MCP server at `https://mcp.particle.pro`, for Claude, Cursor, Codex, and any MCP client. For enterprises, a real-time episode firehose and bulk data snapshots. And pay-per-request over x402 or MPP — call any billable endpoint or MCP tool with no account and no key by paying in USDC on Base.
What gets indexed?
150,000+ podcasts and counting, including the entire Apple Podcasts Top 200 (US) across 130+ English-language verticals plus the long tail of independent shows our customers track. Each show is ingested continuously — new episodes are typically transcribed and enriched within minutes of publication.
How can I search?
Three modes share the same endpoint: an exact keyword or phrase, a semantic match (natural language), or a canonical entity (filter by a person, company, or topic). Each result is a speaker-labeled transcript window with timestamps. Full reference at docs.particle.pro/podcasts/search.
Are transcripts diarized and speaker-identified?
Yes. Every transcript ships with full speaker diarization, and voices are resolved to real people — host, guest, advertiser, correspondent — with role labels and a link to the canonical entity record. Word-level timestamps and resolved entity mentions live on their own endpoints (`/transcript/words` and `/transcript/mentions`) alongside the main transcript.
Can I get audio clips, not just transcripts?
Yes. Each episode's full audio is available, and we also segment every episode into topical clips with their own audio file, transcript window, and engagement scoring. You can carve a custom slice by start/end seconds for any episode, and embed it on a third-party page via Particle's clip widget.
What transcript formats are available?
Three formats from the same transcript endpoint via a `format` parameter: `dialogue` (speaker-attributed lines with timestamps — the default), `text` (plain text), and `srt` (SubRip subtitles). Word-level transcripts live at `/transcript/words` for use cases that need per-token timing.
How quickly are new episodes available?
Most episodes are fully transcribed, diarized, and enriched within minutes of publication. The status of the ingestion pipeline is published at status.particle.pro.
Can I use Particle with my AI agent — Claude, Cursor, Codex?
Yes. Particle ships an MCP server at `https://mcp.particle.pro` that any MCP-compatible client can connect to. Install commands for Claude Code, Claude Desktop, Cursor, and VS Code are at docs.particle.pro/mcp/overview. Agents that hold a wallet and no key can pay per tool call with x402 or MPP instead.
Can I track a specific entity or guest?
Yes. Alerts watch every indexed podcast for a canonical entity — a person, company, brand, or topic — or for a guest appearance, and notify you as-it-happens or in a daily or weekly digest, each match with the clip, transcript, and metadata attached. Delivery is by email, Slack, or webhook. Alerts are entity-based rather than free-text keyword matches, so mentions resolve to the same record wherever they appear.
How does pricing work?
Every plan starts with a $10 free credit (about 1,000 requests). After that, Individual is $29/month for 10,000 requests and Business is $399/month for 100,000, with overages billed per request at list prices (half off on Business). Enterprise tiers and custom contracts are available too — see the pricing page or contact us.
Can I integrate the data into my product?
Yes — the REST API is built for productization. Pull episodes, clips, transcripts, entities, and rankings directly into your platform. Full reference and rate-limit guidance at docs.particle.pro; read the API Terms of Service for the full rules and restrictions.

We value your privacy

We use cookies to understand how you use our platform and to improve your experience. Click "Accept All" to consent, or "Decline non-essential" to opt out of non-essential cookies. Read our Privacy Policy.