Draven AI DocsAPI Reference
Usage & Logs
Every /v1/* request is logged and visible in the console, so you can debug what was sent, by which key, and how it performed.
What's recorded per request
| Field | Description |
|---|---|
| model | The public model id requested. |
| endpoint | The route called, e.g. /v1/chat/completions. |
| status | HTTP status code returned. |
| input / output tokens | From the provider's own usage numbers. |
| tokens_per_second | Computed as output_tokens / (latency_ms / 1000). |
| latency | Time from request received to response returned. |
| key | Which of your API keys was used (shown by its prefix only). |
| app_name / app_url | From the optional X-Title / HTTP-Referer headers below. |
Identifying your app in the logs
Following the same convention as OpenRouter, you can send two optional headers on /v1/chat/completions to label requests from a specific app or project:
Optional headers
X-Title: My Cool App
HTTP-Referer: https://my-app.example.comThese show up as the App / URL column when you filter logs in the console — useful once you have more than one integration sharing the same account.
Viewing logs
Open Logs under Developers → API on dravenai.lat to filter by model, status (2xx/4xx/5xx, or an exact code), endpoint, and date range.
The console Overview page also shows requests and tokens used today against your daily limits, plus how many keys you have and the current model catalog.