Skip to content
OpenRelay is in early access, and the /v1 API is stable. New capabilities ship in the changelog.
Skip to Content
API ReferenceInferenceModel Imports

Get a model artifact

One of your organization's model artifacts: its architecture, dtype, quantization, size, and the hardware it was found to fit or refused. Another organization's artifact reads as not found.

Required permission

Requires the artifact.read permission. See Permissions for the roles and API key permissions that include it.

Command line

The orl CLI wraps this endpoint:

orl imports artifact get <artifact-id>
GET
/v1/inference/artifacts/{artifactId}

Authorization

apiKey
headerAuthorizationBearer <token>

OpenRelay API key. Send it as Authorization: Bearer or_… (legacy vl_… keys still work).

Path Parameters

artifactId*string

Response Body

Artifact

application/json
  1. response
id*string
importId*string

The import that produces the artifact.

status*string
Value in"pending""ready""failed"
weightsRef*string

The weights ref a deployment names to serve it (art:).

digest?string

The digest of its manifest's file set, once ready.

format?string
architecture?string
dtype?string
quantization?string

The format the model's configuration says the weights are quantized to (nvfp4, mxfp4, fp4, fp8, int4, int8), or the method when it says no more; absent when they are not quantized.

paramCount?integer
Formatint64
bytes?integer
Formatint64
files?integer
compat?

The artifact evaluated against one version of a placement policy: the shapes it fits, in policy order, and why each other shape does not.

Example Request
curl -X GET 'https://api.openrelay.inc/v1/inference/artifacts/string'
Example Response
{  "id": "string",  "importId": "string",  "status": "pending",  "weightsRef": "string",  "digest": "string",  "format": "string",  "architecture": "string",  "dtype": "string",  "quantization": "string",  "paramCount": 0,  "bytes": 0,  "files": 0,  "compat": {    "evaluatedAt": "2019-08-24T14:15:22Z",    "policyId": "string",    "policyVersion": 0,    "fits": [      "string"    ],    "refusals": {      "property1": "string",      "property2": "string"    }  }}