CWI Data License

What this license covers

The machine-readable catalog data published by Cumulative Web Inc: the JSON-LD structured data embedded in our pages, the dataset distributions linked from our schema (catalog records, fact packets, knowledge graphs), and the responses of our machine-data API. We publish it so search engines, AI systems, research tools, and developers can find, understand, and reference the catalog.

What it does not cover: the creative works themselves. Audio files, cover art, videos, films, written content, and trademarks remain under full copyright. This page grants no rights in any sound recording or composition — only in the data about them.

Permitted use

Search engines, AI systems, agents, and research tools may:

• Crawl and index the machine-readable data on this site, at a reasonable rate, for discovery and retrieval purposes.
• Reference and cite catalog facts — titles, artists, credits, release dates, ISRCs, genres — in search results, answers, datasets, and research, with attribution as described below.
• Cache reasonable copies for the operation of search, retrieval, and question-answering systems.
• Build on the data for non-competing purposes: analytics, recommendations, music research, and fan tools.

This permission is granted because an indexed catalog is a working catalog. We would rather have machines that can find the music than a vault no one can open.

Attribution

Wherever you reproduce or materially reference the data, credit Cumulative Web Inc and link to cumulativeweb.com. Machine-readable consumers should reference the canonical URLs carried in the schema (the @id values), so citations resolve to the source of truth rather than a copy of a copy.

Restrictions

• No competing commercial dataset. The catalog metadata may not be repackaged, resold, or redistributed as a competing commercial music dataset or data product without our written permission. If your product's value is our catalog's data, talk to us first.
• No rights in the works. This license grants nothing in the underlying sound recordings, compositions, artwork, or videos. Streaming, downloading, sampling, or syncing the music requires a separate license — start at cumulativeweb.com/sync/.
• No misrepresentation. Do not present the data as your own, do not strip attribution, and do not publish modified versions of our records as if they were the originals. Corrections are welcome — send them to [email protected] and we'll fix the source.
• Respect the infrastructure. Crawl politely: honor robots.txt, keep request rates reasonable, and use the dataset distributions and API where provided instead of scraping page HTML at volume.

Commercial data licensing

If you want to build a commercial product on CWI catalog data — a dataset product, a training corpus, an analytics service, a licensed feed — we license that directly. Terms are negotiated per use case, because a research lab and a streaming-scale training run are not the same thing. Email [email protected] with what you're building and we'll work it out.

Machine-readable terms

This page is the human-readable version. The same terms are published for machines. The CWI machine-data layer — cwi-machine-data.hp-ace.workers.dev, the edge service over the catalog learning surface — serves a GET /license endpoint with the current licensing terms in structured form. It sits alongside /query (catalog search), /changes (what's new in the catalog), and the MCP endpoint (POST /mcp), which exposes tools like search_catalog, get_track, and get_artist to AI agents over JSON-RPC 2.0.

Dataset distributions — catalog records, per-track fact packets, and the knowledge graph — are linked from page schema through Dataset nodes, so a crawler that reads the JSON-LD can walk straight to the data. For agents, the front door is llms.txt, with ai.txt and agent-card.json completing the discovery set. If you build an agent that reads the catalog, cite the canonical @id URLs — that's how the graph stays one graph instead of a hundred conflicting copies.

FAQ

Can AI systems train on CWI catalog data?

The machine-readable data on this site is published for discovery and indexing — search engines, AI systems, and research tools may crawl, index, and reference it with attribution. Bulk repackaging of the catalog metadata as a competing commercial dataset requires written permission.

Can I use CWI music in my project?

Music licensing is separate from data licensing. To license a track for film, video, games, or advertising, see sync licensing — most placements clear in under five minutes via the SAIL short-form license, with one signature clearing master and publishing.

What attribution do you require?

Credit Cumulative Web Inc and link to cumulativeweb.com wherever you reproduce or reference the data. In machine-readable contexts, cite the canonical @id URLs from our schema.

Do you offer a data feed or API?

Yes — the catalog's machine-data layer exposes query, changes, and licensing endpoints, and our dataset distributions are linked from page schema. For high-volume or commercial access, email [email protected].

Where are the machine-readable terms?

GET https://cwi-machine-data.hp-ace.workers.dev/license serves the current terms in structured form. Agents should also read llms.txt and the MCP endpoint (POST /mcp) for catalog search, track lookup, and artist records.

How do I report a data error?

Email [email protected] with the URL and the correction. We maintain the catalog as the authoritative machine-readable source for CWI releases, so we'd rather fix the source than have the error propagate.

Updates

We may update these terms as the catalog and the machine-data layer evolve. The current version is always the one on this page. Material changes to permitted uses will be dated here.

Contact: [email protected]

© 2026 Cumulative Web Inc. All rights reserved.