Process

A managed process, with a written record at every step

Rights and quality are separate decisions, and nothing is delivered until both are recorded.

At a glance

Four phases, from request to access

  1. 01Request or submitRequirements from buyers; metadata from suppliers
  2. 02 · decisionQualify & reviewFit check, rights review, technical QA
  3. 03Evaluate a sampleTime-limited, once we approve your request
  4. 04Sign, pay, get accessLicence signed online (we sign first), invoice paid, then read-only access to a private copy

In detail

Every step, and what it produces

  1. Step 01

    Requirements or submission

    Buyers describe the data they need — types, languages, volume, intended use. Suppliers describe the dataset in a short form and connect the GitHub or GitLab repository that holds it, read-only (never credentials or personal data).

  2. Step 02

    Qualification

    We confirm fit and intent, ask clarifying questions, and tell you early if we cannot help. Nothing is promised that we cannot deliver.

  3. Step 03

    Rights review

    We record the legal owner, the submitter’s authority, inbound licences, restrictions, and which uses are permitted. This is a separate decision from technical quality.

  4. Step 04

    Technical QA and preparation

    Structure, secrets, personal data, duplicates, and test validity are checked and reported. Preparation services bring data to the buyer’s schema.

  5. Step 05

    Sample evaluation

    Once we approve a request, the buyer gets a time-limited sample, taken from the reviewed repository, to evaluate.

  6. Step 06

    Licence, payment and access

    Scope, term, exclusivity, and price are set out in a written proposal. ModelData Exchange signs the licence first and the buyer signs last, then pays the invoice by card or bank transfer. Once it’s paid, the buyer’s team gets read-only access to a private copy of the reviewed repository, with its dataset card and rights record.

Start with a conversation.

Share your requirements or submit a dataset — we’ll respond with next steps.