Monitoring API
The Monitoring API exposes GPTfake’s live and historical censorship data: refusal rates, bias scores, and policy-change timelines for every model we track. It is the programmatic surface of the same data shown on our monitoring pages, scored by our published methodology.
Base URL
https://api.gptfake.com/v1All endpoints require a bearer token — see Authentication.
Endpoint summary
| Method & path | Purpose |
|---|---|
GET /monitoring/models | List monitored models — see Models |
GET /monitoring/{model}/metrics | Current metrics — see Metrics |
GET /monitoring/{model}/history | Historical trends |
GET /monitoring/compare | Compare metrics across models |
GET /bias/{model} | Detailed bias analysis |
GET /alerts | List detected changes — see Alerts |
POST /webhooks | Subscribe to events — see Webhooks |
Example
curl https://api.gptfake.com/v1/monitoring/chatgpt/metrics \
-H "Authorization: Bearer YOUR_API_KEY"{
"model": "chatgpt",
"version": "gpt-4o",
"metrics": {
"overall_refusal_rate": 18.7,
"political_refusal_rate": 34.2,
"safety_refusal_rate": 67.8,
"bias_score": -12
},
"trend": "rising",
"last_updated": "2024-11-25T00:00:00Z"
}Every response carries a last_updated timestamp. Figures shown in these docs are illustrative; live values follow the refresh cadence in our methodology. Each metric ships with a sample size so you can weight it.
GPTfake