Data licensing

The data behind the site, without the crawl.

Everything you can see on audioscrape.com is built from one continuously updated dataset: who is mentioned where, how often, alongside whom, and in which recordings. If you want that data in bulk, license it. Crawling the site page by page gets a stale, partial copy and gets blocked.

Request access API reference
What you can license

Four layers, from graph to audio.

Entity knowledge graph

People, organizations, topics and places extracted from spoken audio, with identifiers, aliases, and the recordings each one appears in. Updated nightly.

Mention trends

Mention counts per entity per time window, co-mentions, and the ranked trend tables shown on the public graph, as time series rather than page snapshots.

Podcast & episode metadata

Show and episode records, publication dates, chapters, speakers, and the summaries and highlights generated for each transcribed recording.

Transcripts & search

Full transcripts with speaker attribution and timestamps, and text and semantic search over the corpus, through the API and MCP integrations, on a plan that fits your volume.

Delivery

Bulk exports as JSONL or Parquet on a schedule you set, the REST API with rate limits sized for your use, or webhooks and keyword feeds for changes as they land.

Terms

Pricing follows volume and use, and is quoted per request. Licenses cover redistribution rules, refresh cadence, and attribution. Model training is a separate conversation, not covered by default. Automated bulk access outside the API or a license is prohibited by our Terms.

Why not just crawl it

Because the page is not the dataset.

Public pages show a preview of each recording and a ranked slice of the graph for one time window. The underlying tables are far larger, change every night, and carry identifiers you cannot recover from HTML.

Automated crawling is blocked

Large-scale automated fetching of the site is stopped at the edge, and our robots rules disallow it. Search engines and AI assistants that identify themselves remain welcome.

Previews are previews

Anonymous pages show an excerpt of each transcript by design. A crawl collects excerpts, not transcripts.

Licensing is faster

A bulk export covers the whole corpus in one delivery, with a refresh schedule, instead of weeks of page fetching that produces a copy already out of date.

Request access

Tell us what you are building.

Which layers you need, how much, how often, and what for. We reply with a scope, a price, and a sample, usually within a business day.

sales@audioscrape.com Enterprise options See self-serve plans →