Observability

Every request made with your API key is logged. View your usage, token counts, latency, and cost from the dashboard in real time.

Usage dashboard

Go to Dashboard to see:

Wallet balance

Current credit balance in INR. Updated after every request.

Total requests

Request count over your chosen time range (today, 7d, 30d).

Token usage

Input and output tokens broken down per request.

Estimated cost

Cost per request and total spend in the selected period.

Per-key breakdown

See which API key is generating the most traffic.

Latency

P50 and P95 request latency from your requests to the gateway.

What is logged per request

timestampUTC time the request was received by the gateway.

api_key_prefixThe first 8 characters of the key used (e.g. sk-live-a1b2c3d4).

modelThe model ID passed in the request body.

prompt_tokensNumber of tokens in the input messages.

completion_tokensNumber of tokens generated in the response.

total_tokensSum of prompt and completion tokens.

cost_centsAmount deducted from your wallet in paisa (1 INR = 100 paisa).

latency_msTotal time from request received to first token returned.

statusHTTP status code returned to the caller.

Low balance alerts

Set an alert threshold in Dashboard → Settings. When your wallet drops below that amount, InferexAI sends an email notification and fires a low_balance webhook event (if configured). Alerts are rate-limited to once per 24 hours so you are not flooded.

When the wallet reaches zero, requests fail with HTTP 402. Top up or wait for the next subscription cycle to restore access.

Tip: Check your balance programmatically by calling GET /portal/me with your session token — the response includes wallet_cents.