Steinkauz AI
Steinkauz connects different AI inference providers under one interface you govern:
routing, policy, Budgets, and audit configured to your needs.
Consumers
Inference environments
Public managed
OpenAI
Anthropic
MistralPrivate managed
Amazon Bedrock
Azure OpenAI
Google Vertex AISelf-hosted
Governance
Set policies, allocate Budgets, enforce routing rules, and audit every request before and after inference runs.
Label how sensitive each conversation is and which providers you trust. Steinkauz blocks incompatible requests before inference runs.
Learn moreSearchable activity across chat and API: model steps, tool calls, costs, and API key attribution in one place.
Learn moreChoose how much human oversight each Native Tool needs in chat, from approval before execution to autonomous runs.
Learn moreCap spend per member or API key and per provider. Organizations allocate the System Gateway pool; BYOK stays uncapped until you set optional caps.
Learn moreConsumers
Chat in the browser and API access for software and workflows: two consumers of one control plane, sharing routing, Budgets, and audit.
Chat
Chat, compare, and refine in a native chat UI on the same control plane.
API
Connect OpenAI-compatible clients and your own software to Steinkauz. Inference stays swappable; keys, Budgets, and routing stay yours.
curl -sS "$STEINKAUZ_BASE_URL/v1/chat/completions" \
-H "Authorization: Bearer $STEINKAUZ_API_KEY" \
-d '{"model": "steinkauz-gateway/openai/gpt-5.5", "messages": [{"role": "user", "content": "Hi!"}]}'Sovereignty & Responsibility
Hosting, keys, privacy, and interoperability sit under one control plane, set by you, not inherited from a single vendor stack.
Attach managed Gateway access, your own provider keys, private cloud endpoints, or self-hosted models. How much you operate yourself is a choice on the same dial.
Metadata shows country of origin for AI models and inference providers. Separately, each connection gets an environment class you assign. Routing follows the trust you configure.
Encryption in transit and at rest on Steinkauz servers, plus optional zero data retention on supported providers. No data path you did not choose.
The API is OpenAI-compatible, and routing rules are explicit configuration. You can point software elsewhere without a rewrite: honest, bounded switching cost instead of vendor lock-in.
Deployment
Public cloud for individuals and organizations. Private cloud or on-premises when the control plane has to sit inside your trust boundary. One application, one set of controls.
Self-service
Hosted by us in the EU. Start on your own; when other people join, add an organization with roles, shared policy, and per-seat billing.
Contact us
Steinkauz inside your own environment: private cloud, dedicated tenancy, or on-premises, arranged with our team.
Note
Where Steinkauz runs is independent of where inference runs. You can stay on public cloud and still route sensitive workloads to private or self-hosted providers. See plans and deployment options