Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/GingerlyData247/SOTeam4-P2/llms.txt

Use this file to discover all available pages before exploring further.

​
Endpoint

POST /artifact/model/{id}/license-check
Checks whether a GitHub repository’s license is compatible with a model’s license for fine-tuning or inference purposes. This helps ensure legal compliance when using model weights with custom training or inference code.

​
Path Parameters

​
id
string
required
The unique identifier of the model artifact to check

​
Request Body

​
github_url
string
required
The HTTPS URL of the GitHub repository to check compatibility withFormat: https://github.com/{owner}/{repo}

​
Response

Returns a boolean value indicating compatibility.
​
compatible
boolean
true if the GitHub repository’s license is compatible with the model’s license, false otherwise

​
License Compatibility Matrix

The endpoint uses a predefined compatibility matrix based on standard open-source license compatibility rules:

​
Apache 2.0 Model License

Compatible with:
  • MIT
  • BSD 3-Clause
  • BSD 2-Clause
  • Apache 2.0

​
MIT Model License

Compatible with:
  • MIT
  • BSD 3-Clause
  • BSD 2-Clause

​
BSD 3-Clause Model License

Compatible with:
  • BSD 3-Clause
  • MIT

​
BSD 2-Clause Model License

Compatible with:
  • BSD 2-Clause
  • MIT

​
GPL 3.0 Model License

Compatible with:
  • None (GPL requires derivative works to also be GPL-licensed)

​
CC-BY 4.0 Model License

Compatible with:
  • None (Creative Commons licenses have specific attribution requirements)

​
Fine-Tuning vs Inference

While this endpoint provides a single compatibility check, license implications differ by use case:

​
Fine-Tuning (Training)

When fine-tuning a model:
  • Model weights are modified, creating a derivative work
  • More restrictive license terms apply
  • Training code license must be compatible with model license
  • Output model inherits license constraints from base model

​
Inference Only

When using a model for inference:
  • Model weights are not modified
  • More permissive interpretation may apply
  • Inference code and model can often have different licenses
  • Check specific license terms for production use
Important: This endpoint provides a simplified compatibility check. Always review full license texts and consult legal counsel for production deployments.

​
Example Request

curl -X POST https://api.example.com/artifact/model/42/license-check \
  -H "Content-Type: application/json" \
  -d '{
    "github_url": "https://github.com/huggingface/transformers"
  }'

​
Example Responses

​
Compatible Licenses

true
In this example:
  • Model 42 has an Apache 2.0 license
  • The huggingface/transformers repository uses an Apache 2.0 license
  • These licenses are compatible

​
Incompatible Licenses

false
In this example:
  • Model has a GPL 3.0 license
  • GitHub repository uses an MIT license
  • GPL requires derivative works to also be GPL-licensed
  • These licenses are incompatible for creating derivative works

​
License Detection

The endpoint automatically detects licenses from:

​
Model License

  1. Artifact metadata (cached from ingestion)
  2. Hugging Face API (/api/models/{id} endpoint)
  3. Model card metadata fields

​
GitHub License

  1. GitHub API (/repos/{owner}/{repo}/license endpoint)
  2. Returns SPDX license identifier (e.g., “apache-2.0”, “mit”, “gpl-3.0”)
  3. Automatically parses LICENSE file in repository root

​
Error Responses

​
404
error
Artifact not found: The specified model ID does not exist in the registryORInvalid GitHub URL: URL format is incorrect or does not point to a valid repository
​
502
error
Unable to fetch license: Could not retrieve license information from Hugging Face or GitHub APIsPossible causes:
  • Model or repository does not have a LICENSE file
  • API rate limits exceeded
  • Network connectivity issues
  • Private repository (not accessible without authentication)

​
Behavior Notes

  • Case-insensitive matching: License identifiers are normalized to lowercase
  • SPDX format: Uses standardized SPDX license identifiers (e.g., “apache-2.0” not “Apache License 2.0”)
  • Strict matching: Only explicitly compatible licenses return true
  • Unknown licenses: If either license cannot be determined, the endpoint returns an error
  • External APIs: Requires network access to Hugging Face and GitHub APIs

​
Use Cases

  • Pre-deployment validation: Check license compliance before deploying models in production
  • Training pipeline safety: Verify that training code and model weights have compatible licenses
  • Automated compliance: Integrate into CI/CD pipelines to prevent license violations
  • Repository selection: Filter GitHub repositories by license compatibility with specific models
  • GET /artifact/model/{id}/rate - Returns the numeric license metric score (0.0 to 1.0)
  • GET /artifact/model/{id}/lineage - Shows parent models that may have different licenses

Build docs developers (and LLMs) love

Get started for freeTalk to us