Per-request cost attribution across every provider. See exactly where your tokens go, which models burn the most, and where to optimize.
Grepture tracks token usage and calculates cost for every request flowing through your app. You get per-request, per-model, and per-endpoint cost breakdowns — across OpenAI, Anthropic, Google AI, and every other supported provider. Works in both proxy mode and trace mode, so you get cost visibility even when requests go directly to the provider.
The proxy counts input and output tokens per request and maps them to each provider’s pricing. Costs are attributed to the model and endpoint used. View spend in the dashboard, filter by time range, model, or endpoint, and export for billing. In trace mode, the SDK extracts token counts client-side and sends them async — same cost data, zero proxy overhead.
| Model | Requests | Tokens | Cost | % of total |
|---|---|---|---|---|
| gpt-4o | 892 | 1.2M | $8.94 | 71.7% |
| claude-3.5-sonnet | 412 | 340K | $2.55 | 20.4% |
| gpt-4o-mini | 389 | 890K | $0.54 | 4.3% |
| gemini-1.5-pro | 154 | 210K | $0.44 | 3.5% |
npm install @grepture/sdk — wrap your client — done.
Free for up to 1,000 requests/month · No credit card required
Get Started Free