Machine-readable contracts · fields, sources, samples, discovery

Stable JSON and text routes give buyers, docs tooling, crawlers, and agents one contract surface for field, source, sample, and discovery metadata.

Treat these routes as public read contracts. They describe the current website-facing dataset shape; production data access still uses paid snapshots, REST, MCP, or the authenticated buyer portal.

Contract index

Field catalogue

Field metadata

Schema-backed field groups, descriptions, examples, provenance notes, and per-field documentation URLs.

28 documented fields

/docs/data-dictionary/fields.json

Source catalogue

Source metadata

Schema-backed source codes, jurisdictions, licences, commercial-use flags, source pages, and provenance documentation anchors.

13 verified current-release sources

/sources/catalog.json

Sample manifest

Sample artifacts

Schema-backed evaluation artifact state, row count, release metadata, published files, sizes, checksums, available formats, and source provenance.

150 sample rows · 25 JSON fields · 13 flat columns

/sample/manifest.json

Field catalogue schema

schema/field-catalog.schema.json

Reusable JSON Schema for the public field catalogue contract, including non-empty field groups and concrete descriptions.

28 field entries validate here

/schema/field-catalog.schema.json

Source catalogue schema

schema/source-catalog.schema.json

Reusable JSON Schema for source codes, provenance fields, licence metadata, and commercial-use flags.

13 source entries validate here

/schema/source-catalog.schema.json

Sample manifest schema

Public sample metadata

Reusable JSON Schema for the public sample manifest.

published sample contract

/schema/public-sample-manifest.schema.json

LLM discovery

Discovery index

Compact AI ingestion index linking docs, API, sample, source, and schema surfaces.

3 API examples · 4 use cases

/llms.txt

Sample state

The public sample manifest currently advertises 150 rows · 25 JSON fields · 13 flat columns. Use it to verify file names, formats, sizes, checksums, schema counts, source provenance, licence flags, and release metadata before wiring evaluation jobs.

Agent bootstrap

Agents should start with /llms.txt for compact discovery, then read the field and source catalogues before composing API calls. Human reviewers should use the rendered data dictionary and source licence notes alongside the JSON.

contract discoverybash

curl https://rxatlas.dev/llms.txt
curl https://rxatlas.dev/docs/data-dictionary/fields.json
curl https://rxatlas.dev/sources/catalog.json
curl https://rxatlas.dev/sample/manifest.json
curl https://rxatlas.dev/schema/field-catalog.schema.json
curl https://rxatlas.dev/schema/source-catalog.schema.json
curl https://rxatlas.dev/schema/public-sample-manifest.schema.json

API alignment

The field catalogue describes response fields for /v1/drugs. The source catalogue describes the source codes and provenance anchors returned by records and the source registry endpoint. The sample manifest describes only the public evaluation artifact, not paid buyer files.