kayetanaBuild AI features without handing over data.
Product teams want to try AI quickly, support wants less manual work and compliance wants to know where the data is. kayetana gives you a REST API on your own server, with a flat fee instead of per-token billing.
One key per feature, with a monthly limit.
Create a separate API key for each application, such as your support bot or internal search. Every request is counted per key. When a key goes over its monthly limit, the API returns 429 instead of quietly running on.
- Bearer token authentication
- Requests and tokens per key
- Delete keys at any time
From playground to product.
Test in the playground
Try the model, system prompt and temperature until the result fits.
Copy the code
Grab sample code for Python, TypeScript or curl.
Create a key
Set up an API key for the feature and give it a monthly limit.
Go live
Stream responses with Server-Sent Events and keep an eye on usage.
What fintech teams build with it.
Sort support tickets
Route requests by topic like card, transfer or account, with a confidence score per label.
Search your help center
Rank articles by relevance to the user's question with embed and rerank.
Check copy against guidelines
Compare product copy with internal guidelines and approved wording, with citations.
Function calling
Let the model call tools in your application, such as a status lookup.
Structured output
Generate responses in JSON mode that your backend can process directly.
Your own model
Fine-tune a chat model with LoRA on your own conversation examples.
Predictable costs, visible usage.
kayetana costs a flat fee for the platform, with no billing per token or user. The usage overview shows tokens and requests per day, key and model, plus response time and error rate. Check with your compliance team whether that covers your regulatory obligations.
- Time ranges of 7, 30, 90 or 365 days
- Broken down by key and model
- Logs without prompts or responses
See the API in action.
In the demo, we build a request in the playground, create a key and walk you through usage.
Frequently asked questions
Which endpoints does the API offer?
Under /v2 you get chat, embed, rerank, classify, tokenize, detokenize, transcribe, parse, models, datasets, embed-jobs and finetuning.
Can responses be streamed?
Yes. The chat endpoint streams with Server-Sent Events, including answers from documents with citations.
How does pricing work?
A flat fee for the platform. There's no billing per token or user. Pricing is available on request.
Does kayetana store user input from our app?
No. Prompts and responses are never stored. Only time, endpoint, model, status, tokens and duration are logged.
Does kayetana meet financial industry regulations?
We can't promise that. Data stays on your server. Review the requirements with your compliance team and data protection officer.