HELP & SUPPORT
Frequently asked questions
Everything you need to know about KEYVERA. Can't find what you're looking for? Contact us.
Frequently asked questions
What is a unified AI API?
A unified AI API lets you send requests through one consistent interface instead of integrating separately with each model provider's own API.
Which AI providers can I access through KEYVERA?
Keyvera supports models from OpenAI, Anthropic, DeepSeek, xAI, and Zhipu AI. See the Models and Pricing pages for the current catalog and availability.
Can I switch models without changing my integration?
KEYVERA's routing layer is designed so switching models typically requires changing a configuration value rather than rewriting your integration code.
How does model routing work?
Requests can be routed according to configurable policies, such as latency, cost, or capability preferences, instead of being hardcoded to a single provider.
Can I configure fallback models?
Resilient Model Access is designed to support provider fallback and model failover options for applications that need continuity if a provider is unavailable.
Does KEYVERA store prompts or model responses?
Data retention and logging policies are still being finalized. This answer will be updated with confirmed details before launch.
Is Keyvera suitable for production workloads?
Production suitability depends on your required models, rate limits, availability, and data-handling needs. Review the documentation and service terms before deploying critical workloads.