Centrally provided models
Administrators provide a shared model catalog and decide which models may be used. Employees choose from the same model picker.
Models and providers
Model freedom. Centrally managed.
Choose the right model for the task, not a new knowledge platform for every provider. Managed models, BYOK and your own inference share the same AI Workspace.
Model providers
ChatGPT
Claude
Gemini
Mistral AI
ChatGPTClaudeGeminiMistral AIGoverned contextOne policy boundary
Interfaces
Search
AI agents
MCP & API
Local models
Your own keys – BYOK
Connect your own provider accounts. Your organization decides whether personal API keys are allowed or centrally managed connections should be used.
Your own inference
Connect Ollama or vLLM in your infrastructure. Your knowledge layer stays independent of where you run the model.
BYOK
Your providers. One workspace.
Connect a provider, test access and approve models. Model-dependent capabilities such as image inputs remain part of the selection.
OpenRouter
Provider access with your own API key
OpenAI
Provider access with your own API key
Anthropic
Provider access with your own API key
Gemini
Provider access with your own API key
Ollama
Your own inference
vLLM
Your own inference
Operating model
Plan data regions and model choice together.
Align model providers, processing regions, retention and contractual requirements with your deployment. Centrally managed access and BYOK keep their respective billing arrangements.