Data access · Private connections

Deliver datasets into your environment.

A private connection sends licensed marketplace data straight to your warehouse, object storage or application — on a schedule, with provenance and license terms attached — instead of calling the public API.

Talk to the marketplace team Downloads and rights

Status

Rights-gated, like downloads

A private connection is a bulk delivery, so it follows the same rule as downloads: it is only offered where the dataset's license allows redistribution and the provider has confirmed it. The check is enforced in code, not policy: a dataset whose redistribution is restricted or unknown, or whose download channel is awaiting rights confirmation, cannot be delivered, and the request fails with a rights_restricted error rather than a partial file.

None of the current LAW.co datasets can be delivered this way yet — their download channel is awaiting rights confirmation under the pending provider agreement. You can scope a connection now so it is ready when rights are confirmed.

Delivery options

Where data can land

Target	Delivery	Notes
Cloud warehouse	Incremental table sync	Snowflake, BigQuery, Redshift, Postgres
Object storage	Partitioned files	Parquet, JSONL or CSV where the dataset supports the format
Application	Webhook or queue	Change notifications plus a pull endpoint
Data.co Control	Governed catalog	Policies, lineage and audit for the delivered data

Setup

How a connection is set up

1
Confirm rights
We check the dataset's license and download status for your intended use. If redistribution is not confirmed, the connection waits; nothing is delivered on assumption.
2
Choose the target and format
Pick a warehouse, a storage bucket or an application endpoint, and a format the dataset supports. You grant Data.co write access to a dedicated schema, bucket prefix or endpoint — never broad credentials.
3
Map the schema
Dataset fields are mapped to target columns. Provenance and license fields are delivered as columns of their own, and derived fields stay separate from source fields.
4
Initial load
The licensed dataset version is delivered in full, with a delivery log entry recording the version, row counts and checksum.
5
Incremental sync
New and changed records are delivered on the agreed schedule. Schema changes are announced before they arrive, and removals are delivered as explicit deletes.

Delivered table

Provenance travels as columns

An illustrative warehouse table for a delivered legal dataset. Business columns sit next to provenance and license columns, so a query or a dashboard can always show where a row came from and under which terms.

schema.sql

create table dataco.court_opinions (
  id                text primary key,   -- lawco:case:<id>
  title             text,
  citations         text[],
  court             text,
  jurisdiction      text,
  decision_date     date,
  -- provenance
  provider          text,               -- "lawco"
  source_id         text,
  upstream_source   text,               -- e.g. CAP, courtlistener
  canonical_url     text,
  retrieved_at      timestamptz,
  license_id        text,
  is_derived        boolean,            -- false for source text and facts
  -- delivery
  dataset_version   text,
  delivered_at      timestamptz
);

Illustrative example · final API surface documented in /docs

Every delivery includes

Per-record provenance
Dataset version and delivery timestamp
License id and attribution text
Delivery log for reconciliation
Schema change notices
Revocation if rights change
Derived fields kept in separate columns

FAQ

Questions buyers ask

Is this different from a download?

A download is a one-off file. A private connection keeps your copy in sync as the dataset updates and logs every delivery.

Which datasets support private connections today?

None yet. Each LAW.co dataset's bulk channel is awaiting rights confirmation; the dataset pages show the current status.

What happens if a provider's rights change?

Deliveries for that dataset stop. Depending on the license, you may be required to delete the delivered copy, and the delivery log records the revocation.

Is a private connection metered?

Deliveries are recorded against your organization like downloads. Commercial terms for ongoing delivery are set per dataset and per agreement, because bulk use is usually priced differently from per-request access.

Can we use delivered data to train models?

Only if the provider's license says so. LAW.co's current terms prohibit training on its corpus without a written agreement.

Should we use the API or a private connection?

Use the API for request-time lookups and agents. Use a private connection when you need the data in your warehouse for joins, analytics or a local index — and the license allows a copy.

Which credentials do you need in our environment?

Write access to a dedicated schema, bucket prefix or endpoint only. We do not need read access to the rest of your environment, and you can revoke the credential at any time.

Can deliveries be paused?

Yes. Pausing stops new deliveries; resuming catches up from the last logged delivery.

Dataset Downloads

Parquet, CSV, JSONL where licensed

Dataset Integration

Into your stack and tools

Private Data Deployment

Datasets inside your network

REST APIs

Versioned, paginated, metered