Frequently asked questions
Does LEANOX.AI change my model's responses?
No. Your request goes to the same provider and model, and you get the provider's response in the provider's format. LEANOX.AI never changes what your prompts say. It changes how their repeated parts are cached and, with an opt-in setting coming soon, the order in which repeated blocks are sent, so that more of them can be served from the cache.
Which providers can I use?
Anthropic, Google Gemini (with an API key, or via Gemini Enterprise Agent Platform), OpenAI and xAI. Savings are available today for Anthropic and Gemini. For OpenAI and xAI they're coming soon: your requests already go through LEANOX.AI with full Logs and Analytics, and savings will start automatically once they're available.
Who bills me for model usage?
Your provider, directly, as before. Your provider bill is the final price after LEANOX.AI savings are applied.
Why doesn't Estimated Spend match my provider invoice exactly?
Spend is calculated from published list prices. Discounts, credits, taxes and price changes on your provider account can make the invoice differ.
Is my LEANOX.AI quota the same as my provider's quota?
No. Your LEANOX.AI quota counts the tokens your requests use through LEANOX.AI each month. Your provider's own quotas and rate limits still apply on your provider account, separately.
I lost my API key. Can I see it again?
No — keys are shown only once. Create a new key, update your application, then revoke the old one.
Can other users see my data?
No. Analytics and Logs show only your own requests.
Troubleshooting
| Symptom | Likely cause and fix |
|---|---|
| Authentication error (401) | The LEANOX.AI key is wrong, incomplete, or revoked. Check that the key starts with aip_sk_ and is active on the API Keys page. |
| Requests to one provider fail | No provider key is stored, or it was removed or has expired at the provider. Check the provider's row on the Providers page and use Rotate to update it. |
| Gemini Enterprise Agent Platform requests fail | Make sure the provider is enabled, vertexai=True is set, and no key is passed in api_key or GOOGLE_API_KEY (see Send your first request). |
| All requests blocked, red banner showing | Your monthly LEANOX.AI token quota is used up. Requests resume on the date shown, or you can contact us to ask about paid upgrades. |
| A new request doesn't show up yet | Requests usually reach Logs and Analytics within a few minutes. Try reloading the page in another minute or two. In Analytics, check that the key filter is set to All or to the key you used. |
| Wrong URL errors (404) | Check the base URL against the table in Send your first request. |
Glossary
- API key (LEANOX.AI key) — the
aip_sk_key your application uses to call LEANOX.AI. - Provider token — your own provider key (Anthropic, Gemini, OpenAI or xAI), stored in LEANOX.AI so it can call the provider for you.
- Cached tokens — input tokens the provider served from its cache, which are billed at a lower rate.
- Estimated Spend — our estimate of the cost of your requests based on the provider's list prices.
- Savings — how much less you paid thanks to LEANOX.AI.
- Quota — the number of tokens your LEANOX.AI plan includes each month. It's separate from your provider's own limits.
Support
Questions or problems? Contact us, or click Feedback at the top of any dashboard page. If it's about a specific request, include its Request ID from Logs.