Skip to content
tiyuvta
  • Research
  • Blog
  • Inference
  • Docs
  • About
  • memra
  • Research
  • Blog
  • Inference
  • Docs
  • About
  • memra on GitHub

status

These terms are a draft pending owner sign-off (spec §12.8). They are published for review and are not yet in force.

Terms of service

draft · 2026-08-05

1. Who you are contracting with

The tiyuvta inference API and this website are operated by [LEGAL ENTITY — owner sign-off §12.8]. By using the service you agree to these terms.

2. The service, as it is

tiyuvta provides an inference API serving open-source models. At launch the service is provided as-is: there is no service-level agreement and no uptime guarantee. Measured reliability numbers are published as receipts on this site; they describe what has happened, not what is promised. This mirrors what the About page says plainly: no SLA is sold until one has been earned.

3. Rate limits and abuse

Requests are subject to rate limits, communicated truthfully via response headers. Rate-limited requests receive 429 + Retry-After — that behavior is the product, not a fault.

Prohibited: attempting to evade rate limits or metering, resource abuse (including automated signup abuse of evaluation credits), using the service to generate content that is unlawful in your jurisdiction or [GOVERNING JURISDICTION — owner sign-off §12.8], and attacking the service or other tenants. Accounts engaged in abuse may be limited or terminated; where practical you will be told why.

4. Your data

Data handling is governed by the privacy & data policy. In short: request metadata is logged for metering; prompt and completion bodies are not logged; your data is never used for training.

5. Billing

Usage is metered per token as published on the pricing page. The meter is honest: cached tokens are itemized, and aborted requests are billed to the abort point. Disputed charges: email [email protected] with the request ids; the metering logs decide.

6. Liability

To the maximum extent permitted by law, the service is provided without warranties of any kind, and liability is capped at the amount you paid for the service in the preceding [LIABILITY PERIOD — owner sign-off §12.8].

7. Changes and termination

These terms may change; changes are announced with dates before they take effect. You may stop using the service at any time; unused prepaid balances are handled per the pricing page. Model ids are pinned and deprecations are announced with dates.

8. Governing law

These terms are governed by the laws of [GOVERNING JURISDICTION — owner sign-off §12.8], and disputes are subject to the courts of that jurisdiction.


tiyuvta

An AI research lab. Every claim on this site links to a raw log in the public memra repo — the result can be yes, no, or not yet.

© 2026 tiyuvta

lab
  • Research
  • Blog
  • About
  • RSS
inference
  • Overview
  • Models
  • Pricing
  • Docs
  • Status page — coming soon
elsewhere
  • GitHub
  • avifenesh.ai
  • Privacy
  • Terms
  • Contact