About PodLume

We turn what is said on podcasts into a map you can search.

Podcasts are where founders, investors, experts and politicians talk at length, often before anyone writes it down. Almost none of it is searchable. PodLume listens to hundreds of top shows, records who said what, and links it all together so you can find it, measure it and check it.

What is underneath

A knowledge graph built from audio, that keeps time

Each episode is transcribed and each speaker identified. Every person, company, product, idea and claim becomes one record, linked to the shows, speakers and moments where it came up.

Because every link carries a date, the graph knows what changed this week compared with usual. That is how we spot a trend early, and why every trend can be traced back to the second it was said.

CoreWeave

Example record

One record · company

Discussed on
35 shows in one week
Said by
Hosts, analysts and investors
Claims
“A financing issue, not weak demand”
Discussed with
Nebius, AI data centers

Over time

Its busiest week: 14× its usual share of the conversation

An example with simplified figures, not live data. See CoreWeave's live record

What we build on it

Four ways into the same map, all on one account.

  • Signals

    What market experts are calling on stocks, sectors and events, counted one expert at a time. For investors.

    Read the latest briefing
  • Trends

    Which names are rising, how they sound and why, with every claim cited. For comms, marketing and research.

    Explore trends
  • Advertiser intelligence

    Who sponsors which shows, who is new and where the reads land. For brands, agencies and podcast sales.

    Browse advertisers
  • Smart Player

    Your own podcasts with transcripts, highlights and answers you can trace. For listeners, on the web and iPhone.

    See the Smart Player

How we work

Every claim cited
Summaries and statements link to the episode and the second they come from.
Estimates labelled
Where a number is modelled, such as reach, we say so and explain how.
Method in the open
How we choose shows, count mentions and leave out ad reads is in How it works.
Corrections welcome
See a misheard name or a duplicate? Tell us and we fix the record.

Want the graph for your own sources?

If you need this kind of map for other audio or text, or access through an API, tell us what you need.

Contact us
About · PodLume