# Check connectivity to an OpenAI-compatible VLM endpoint

Endpoint: POST /probes/vlm-endpoint
Version: 2.0.0
Security: engineBasic

## Request fields (application/json):

  - `url` (string,null)
    OpenAI-compatible base URL; absent or null uses the reference's effective URL, and then its stored key unless an explicit api_key is given. When present, the reference's stored key is not used.

  - `api_key` (string,null)
    Explicit key to probe with; wins over any reference's stored key.

  - `stream_group_config` (string,null)
    A stored stream group config by name, whose effective VLM endpoint is the probe target.

  - `application` (string,null)
    Engine application of the stream to probe, paired with stream.

  - `stream` (string,null)
    Name of the running stream whose effective VLM endpoint is the probe target.

## Response 200 fields (application/json):

  - `reachable` (boolean, required)
    Whether the endpoint answered.

  - `models` (array,null)
    Model names the endpoint reports; null when it does not answer or offers none.

  - `error` (string,null)
    Why the probe failed, when it did.

## Response default fields (application/problem+json):

  - `type` (string)
    URI identifying the problem type.

  - `title` (string, required)
    Short, human-readable summary of the problem type.

  - `status` (integer, required)
    The HTTP status code, repeated from the response line.

  - `detail` (string)
    Human-readable explanation specific to this occurrence.

  - `instance` (string)
    URI identifying this specific occurrence of the problem.