Minimal Example

This example demonstrates the minimum required fields for a valid LLM Data Card. It’s ideal for simple datasets that don’t need extensive metadata.

When to Use This Pattern

Use this minimal structure when:

Your dataset is straightforward and publicly available
You want to quickly publish a data card
Your dataset has no personal data or complex licensing

Complete Example

{
  "schema_version": "llm-datacard/v1.0",
  "core": {
    "id": "example-minimal",
    "version": "1.0.0",
    "title": "Example Minimal Dataset",
    "summary": "A minimal example dataset demonstrating the required fields of the LLM Data Card schema.",
    "maintainer": "DataPass Team",
    "contact": "data@meetkai.ai"
  },
  "data": {
    "kind": "real",
    "modalities": ["text"],
    "languages": ["en"],
    "size": {
      "examples": 1000
    },
    "domains": ["general"],
    "record_format": "json-structured"
  },
  "rights": {
    "license": "CC0-1.0",
    "allows_commercial_use": true,
    "contains_personal_data": "none"
  },
  "provenance": {
    "source_types": ["official-open-data"]
  },
  "access": {
    "availability": "public-download",
    "url": "https://datapass.meetkai.ai/registry/example-minimal/1.0.0"
  }
}

Section Breakdown

Core Section

Every data card must include these core identifiers:

Field	Purpose	Example Value
`id`	Unique dataset identifier	`example-minimal`
`version`	Semantic version	`1.0.0`
`title`	Human-readable name	`Example Minimal Dataset`
`summary`	Brief description	1-3 sentences
`maintainer`	Organization or person	`DataPass Team`
`contact`	Email for inquiries	`data@meetkai.ai`

Data Section

Describes what the dataset contains:

Field	Purpose	This Example
`kind`	Real, synthetic, or hybrid	`real`
`modalities`	Types of content	`["text"]`
`languages`	BCP-47 language codes	`["en"]`
`size.examples`	Number of records	`1000`
`domains`	Subject areas	`["general"]`
`record_format`	Record format	`json-structured`

Rights Section

Specifies licensing and data privacy:

Field	Purpose	This Example
`license`	SPDX license identifier	`CC0-1.0`
`allows_commercial_use`	Commercial usage allowed	`true`
`contains_personal_data`	Privacy level	`none`

Provenance Section

Documents data origins:

Field	Purpose	This Example
`source_types`	How data was collected	`["official-open-data"]`

Access Section

Tells users how to get the data:

Field	Purpose	This Example
`availability`	Access method	`public-download`
`url`	Download location	Direct URL

Optional Enhancements

Consider adding these fields as your dataset matures:

core.doi - Digital Object Identifier for citation
core.preferred_citation - How to cite the dataset
data.size.tokens - Token count for text datasets
provenance.collection_start_date / collection_end_date - Collection timeframe
use.intended_uses - Recommended applications

Try It

Open in Builder - Create your own minimal card
Validate this example - Check schema compliance
View in Registry - See the full registry entry