# Luspend

> Luspend is an LLM FinOps proxy for OpenAI and Anthropic. Platform and infrastructure teams in Europe point existing SDKs at one URL, keep provider keys, and track spend by team and feature. Prompts are not stored.

Site: https://luspend.com
Dashboard: https://app.luspend.com
Proxy: https://api.luspend.com

## Product

- Drop-in baseURL change for OpenAI (`https://api.luspend.com/v1`) and Anthropic (`https://api.luspend.com`)
- Organization proxy token on header `x-proxy-token`
- Optional spend tags: `x-proxy-tag-team`, `x-proxy-tag-feature`
- Provider API keys stay in the customer application
- Prompt and completion bodies are not written to Luspend storage
- Dashboard hosted on Vercel in Frankfurt (fra1)
- You still pay OpenAI and Anthropic for tokens; Luspend meters proxy requests

## Pricing (USD)

- Free: 150 requests per calendar month, hard cap
- Solo: $19/mo, 25,000 requests, then $0.50 per 1,000
- Team: $49/mo, 150,000 requests, then $0.25 per 1,000

## Latency (Luspend test, streamed ~140-token reply)

- Direct time to first token: 554 ms
- Via Luspend: 738 ms
- Median extra hop: 184 ms

## Not Luspend

- Not a token reseller
- Not a prompt observability warehouse
- Not a replacement for self-hosted LiteLLM if you need a multi-provider mesh you operate

## Pages

- [Home](https://luspend.com): product, setup, pricing, FAQ
- [Privacy](https://luspend.com/privacy): what is stored