Browse by Topic
- Getting started: base URLs, keys, finding models.
- Models: model IDs, Fusion, supported inputs.
- Privacy and attestation: privacy levels, verifying evidence, retention.
- Usage and billing: keys, spend caps, failed requests, funding.
- Errors and limits: 401, 400, 429, size caps, what is not available.
- Integrations: SDK drop-ins, frameworks, editors, MCP.
Top Questions
Is Nito a model, or a gateway? A gateway. Nito sits in front of many models from several providers, and you choose the model per request. See How the API Works. Do I need a special SDK? No. Point the OpenAI SDK (or plain HTTP) athttps://app.trynito.ai/v1 and use a Nito catalog ID. Anthropic SDKs use https://app.trynito.ai and /v1/messages. See Integrations.
Does every call run in attested hardware?
No. Confidential is one of four privacy levels, selected by using a model’s :confidential variant. Frontier models run at Anonymous and open-weight models at Private. See Privacy levels.
Does Nito store my prompts?
No. Nito stores nothing you send or receive, at any level. See Data retention and visibility.