Paste your OpenAI API key. aiusage routes GPT calls through the same caching + fallback layer that kills your Claude bill. One balance, every provider. Your keys stay yours — we just sweep them under one roof.
Paste your OpenAI key on the dashboard — encrypted the moment we receive it.
Set OPENAI_BASE_URL=aiusage.ai in your SDK. That's it.
GPT-4, GPT-5, o1, o3 — all route through our cache + cost-optimizer. Same prompts, 40–70% cheaper.
Billed from your aiusage runs balance. OpenAI's key pays the upstream; we give you the cache + routing margin back.
Most teams already pay for Claude AND ChatGPT. Without aiusage you have two dashboards, two billing cycles, zero shared cache. Route both through us and your team's repeat prompts hit the same cache — savings compound across providers.
| Usage | Direct OpenAI | Through aiusage | You save |
|---|---|---|---|
| 1K GPT-4 calls/mo | ~$80 | ~$35 | ~$45 |
| 10K GPT-4 calls/mo | ~$400 | ~$160 | ~$240 |
| 50K mixed Claude+GPT | ~$2,000 | ~$700 | ~$1,300 |
No separate per-provider pricing. One runs balance, any model, shared cache.