Documentation

REST API Reference

Integrate the EnvForage dependency matrix resolution logic directly into your custom CI/CD pipelines, DevOps tooling, or setup workflows.

POST/api/v1/compatibility/resolve(Resolves CUDA, Python, and package targets)

Live Interactive Endpoint Simulator

Operating System

ML Framework

Nvidia Driver Version

Simulated Request

curl -X POST "http://localhost:8000/api/v1/compatibility/resolve" \
  -H "Content-Type: application/json" \
  -d '{
    "os": "linux",
    "gpu_vendor": "nvidia",
    "driver_version": "535.104",
    "framework": "pytorch"
  }'

JSON Response (200 OK)

{
  "status": "success",
  "python_version": "3.10",
  "cuda_version": "12.1",
  "packages": [
    "torch==2.1.0+cu121",
    "torchvision==0.16.0+cu121",
    "torchaudio==2.1.0+cu121"
  ],
  "verification_status": "verified"
}

Request Body Parameters

Parameter	Type	Required	Description
os	string	Yes	Target operating system. Allowed: linux, windows, macos
gpu_vendor	string	Yes	Graphics processor driver set. E.g. nvidia, amd, cpu
driver_version	string	Yes	Local graphic driver version identifier. E.g. "535.104", "525.60"
framework	string	Yes	Desired ML framework stack. Allowed: pytorch, tensorflow, jax

HTTP Response Codes

Code	Status	Description
200	OK	Dependencies resolved successfully. Returns compatible packages configuration.
400	Bad Request	Query attributes could not be resolved or driver specifications are missing.
422	Validation Error	Request payload contains invalid properties or fails data scheme validations.
500	Internal Error	Resolution server experienced an issue or cannot query dependency db.