Frequently Asked Questions (FAQ)
Cost & Billing
Section titled “Cost & Billing”How does billing work? Billing is handled via AI-Punkte, token-based on a pay-as-you-grow principle. You only pay for tokens actually processed.
What does a token cost? The AI-Punkte per model are listed in the AI-Punkte calculator. Input and output tokens are billed at different rates.
Are there precommit discounts? Yes. With a commitment, you get higher rate limits and better terms.
Are there fixed costs? No. Billing is purely usage-based, no CAPEX.
Models
Section titled “Models”Which models are available? See the model overview.
Can I bring my own models? For custom models, see our Managed LLM product.
How often are models updated? Models are continuously evaluated; updates happen as needed.
What happens on deprecation? See the Deprecation Policy. Notice periods depend on tier.
Technology
Section titled “Technology”Is the API OpenAI-compatible? Yes, 100% compatible, a drop-in replacement.
Which SDKs are supported? OpenAI SDK, Anthropic SDK, LangChain, LlamaIndex, and LiteLLM.
Are there rate limits? Yes, RPM and TPM. See Rate Limits.
How can I increase my rate limits? Increase your commitment or contact your noris representative.
What is prefix caching? See Prefix Caching.
Do you support streaming?
Yes, SSE streaming via stream: true.
Compliance
Section titled “Compliance”Is my data stored? No. noris operates zero-data-retention.
Is my data used for training? No, under no circumstances.
Where are the servers located? In Germany, in highly secure noris data centers.
Is this GDPR-compliant? Yes. All data stays within Germany/the EU.
Support
Section titled “Support”How do I get an API key? Through your contact person at noris or anfrage@noris.de.
How do I reach support? Via the customer portal servicenow.noris.net.
Is there an SLA? Yes, 99.9% annual availability for the API endpoints.
Can I get a test account? Yes, contact anfrage@noris.de.