REST Endpoints Inventory
REST Endpoints Inventory
The Supervised AI testing API provides a programmatic interface to manage datasets, initiate evaluation runs, and retrieve performance metrics for AI models. All requests must be sent over HTTPS and authenticated using your platform API key.
Base URL
All API requests are relative to the following base URL:
https://api.supervised.ai/v1/testing
Authentication
Requests require a Bearer Token in the Authorization header.
Authorization: Bearer <YOUR_API_KEY>
Evaluation Management
Endpoints used to trigger and monitor AI model evaluations against specified test suites.
Start Evaluation Run
Initiates a new testing session for a specific model against a dataset.
-
Endpoint:
POST /evaluations/run -
Request Body: | Field | Type | Required | Description | | :--- | :--- | :--- | :--- | |
model_id|string| Yes | The unique identifier of the AI model to test. | |dataset_id|string| Yes | The ID of the test dataset to use. | |config|object| No | Optional parameters liketemperatureormax_tokens. | -
Success Response (202 Accepted):
{
"run_id": "eval_88234-x9",
"status": "queued",
"estimated_completion": "2023-11-01T14:30:00Z"
}
Get Evaluation Status
Retrieves the real-time progress of a specific evaluation run.
-
Endpoint:
GET /evaluations/{run_id}/status -
Path Parameters:
run_id(string): The ID returned when the run was initiated.
-
Success Response (200 OK):
{
"run_id": "eval_88234-x9",
"status": "processing",
"progress": 65,
"completed_samples": 130,
"total_samples": 200
}
Dataset Management
Endpoints for managing the ground-truth data used during the testing process.
List Datasets
Returns a paginated list of all test datasets available in the workspace.
-
Endpoint:
GET /datasets -
Query Parameters:
limit(integer): Number of items to return (default: 20).offset(integer): Pagination offset.
-
Success Response (200 OK):
{
"datasets": [
{
"id": "ds_qa_01",
"name": "General QA Benchmark",
"sample_count": 500,
"created_at": "2023-10-15T09:00:00Z"
}
]
}
Upload Test Dataset
Uploads a new set of test cases (JSON or CSV format).
- Endpoint:
POST /datasets/upload - Content-Type:
multipart/form-data - Request Params:
file: The dataset file.name: (string) The display name for the dataset.
Results and Analytics
Endpoints to retrieve the output of completed tests and performance benchmarks.
Get Evaluation Results
Fetches the detailed performance metrics (Accuracy, Precision, Recall, Latency) for a completed run.
- Endpoint:
GET /results/{run_id} - Success Response (200 OK):
{
"run_id": "eval_88234-x9",
"metrics": {
"accuracy": 0.94,
"f1_score": 0.92,
"avg_latency_ms": 450
},
"summary": "Model performed above threshold for 90% of samples."
}
Export Results
Generates a downloadable report of the test results.
- Endpoint:
GET /results/{run_id}/export - Query Parameters:
format:csvorpdf.
- Success Response (200 OK): Returns a binary file stream.
Internal Utilities
Note: These endpoints are intended for internal platform synchronization and may be subject to stricter rate limits.
System Health
Checks the connectivity of the testing engine.
- Endpoint:
GET /health - Role: Monitoring and load balancing.
Internal Cache Purge
Clears temporary evaluation artifacts.
- Endpoint:
POST /internal/cache/clear - Role: Maintenance of storage resources between large test batches.