Three ways to publish a dataset.
Connect your existing API, let us front your corpus with an MCP connector, or deliver files. Whichever you choose, the listing describes rights, provenance, coverage and schema before anything goes live.
Pick the one that matches your data
Connect an API
We proxy requests to your API at query time. Your data stays in your systems; we store references and usage events, not your corpus.
Connector over your corpus
We build a connector to a read-only service interface you expose, as we are doing with LAW.co.
Upload files
Versioned file deliveries for datasets that change on a schedule. Only for data you can license for redistribution.
How the modes compare
| Connect an API | Connector over a corpus | Upload files | |
|---|---|---|---|
| Where data lives | Your systems | Your systems | Data.co storage |
| Freshness | Live at request time | Live at request time | As often as you deliver |
| Work on your side | A service credential | A small read-only interface | A delivery job |
| Bulk download | Only if you offer an export | Only if you offer an export | Possible, if licensed |
| Best for | Mature APIs | Large corpora without a public API | Periodic reference data |
From application to listing
- 1
Apply
Tell us what you hold, where it comes from and your rights to it. A short description and a sample are enough to start.
- 2
Rights review
We confirm redistribution rights, attribution and any third-party terms, layer by layer. Source text, facts and your own editorial or AI output often carry different rights, and each is assessed separately.
- 3
Describe
Fill in the catalog descriptor: coverage, schema, formats, update frequency, quality notes. Counts and date ranges stay empty until they are verified.
- 4
Connect
Issue a dedicated, read-only service credential, or set up file delivery. We build or configure the connector and map your records to the provider-neutral shape with provenance.
- 5
Test
We run real requests against your system: search, single-record fetch, error cases and rate limits. Each access channel gets a status based on what actually works.
- 6
List
Publish with honest statuses per access method. Channels that are not ready stay visible as awaiting credentials or awaiting rights confirmation.
How LAW.co is being onboarded
LAW.co's catalog lists five datasets. Each access method has an explicit status: the public web pages are available; REST and MCP went live once LAW.co issued a dedicated service credential, except the statute reference table, which LAW.co has not yet populated; bulk download is awaiting rights confirmation. One dataset built on third-party LegiScan data is listed as restricted.
Discovery found that LAW.co's existing search endpoint samples results at random and cannot paginate, and that it has no citation-resolution endpoint. Rather than hide that, the connector labels each limit: searches carry a warning, and citation resolution returns a structured unsupported_operation error. A small read-only service namespace with deterministic pagination has been proposed to LAW.co.
The rights review split the corpus into layers: public-domain opinion text and case facts, LAW.co's own AI summaries and citation graph, and third-party material that is not exposed at all. Each layer has its own license id and pricing tier.
That is the standard we hold every provider to: list what exists, show what is blocked and why, and never present an unavailable method as working.
Have these ready
- A description of each dataset and its upstream sources
- The license or ownership basis for each layer
- Any third-party terms embedded in the data
- A sample of records with your current metadata
- API documentation or an export format description
- A technical contact and a rights contact
Questions buyers ask
Do I have to give Data.co a copy of my data?
No. With API or connector publishing, Data.co calls your systems at request time and stores only references and usage events.
Can I publish data I licensed from someone else?
Only if your license allows redistribution. Third-party data without confirmed resale rights is listed as restricted or not listed.
How long does onboarding take?
It depends mostly on the rights review and on how much connector work is needed. We give you an estimate after reviewing your application and sample.
Can I publish before my API is ready?
Yes. A dataset can be listed with its channels marked awaiting credentials, so buyers can discover it while you finish the integration.
Will buyers' agents be able to use my data?
Yes, through the MCP server, which runs the same operations as the REST API with the same rights checks, quotas and metering.
Who handles billing?
Data.co meters usage and bills buyers. Settlement with you is set in a written provider agreement.
List your data on Data.co.
We review every provider for licensing, provenance and quality. You keep control of terms and pricing.