Skip to main content
Models and providers are managed in Command Center under Settings, in two tabs: Providers and Models. The catalog is scoped to your tenant and is available to every organization and agent in it. Adding a provider or model is pure configuration in the UI, with no engine restart and no editing of environment files.
Viewing the catalog is open to any tenant user, but only users with tenant-level access can add, edit, or delete providers and models. See Users & Organizations.

​
Providers

Open Settings > Providers and click Add Provider.
FieldRequiredDescription
KeyYesShort identifier, unique within your tenant. Example: openai, cloudtemple
Provider typeYesThe integration used to reach the provider — see Provider types
Base URLCompatible & Azure typesEndpoint root, so you can point 2501 at any compatible endpoint. Required for OpenAI/Anthropic-compatible and Azure; leave empty for native types
API Key Var NameWhen auth is requiredThe name 2501 resolves to find the secret — see API key resolution
Requires auth—On by default. Turn off for endpoints that need no key (for example a self-hosted model on a private network)
Enabled—Disabling a provider makes all of its models unavailable without deleting anything

​
Provider types

FamilyTypesNotes
Nativeopenai, anthropic, mistral, deepseek, cohere, openrouter, togetheraiBuilt-in integrations. No endpoint URL needed.
Compatibleopenai-compatible, anthropic-compatible, azurePoint at any API-compatible endpoint via a Base URL.
Use a compatible type to register a self-hosted model server (vLLM, Ollama, LM Studio) or any vendor that exposes an OpenAI- or Anthropic-style API.
Adding a new endpoint of an existing type is pure configuration. Adding a brand-new type (a vendor with its own API shape) requires a 2501 release—contact your Account Executive.

​
API key resolution

The API key is resolved at runtime in two steps:
  1. First, 2501 looks for a tenant-scoped Credential with the matching name.
  2. If none exists, it falls back to an environment variable of the same name on the engine.
This lets you rotate keys in the UI without touching the engine environment. When you save the provider, the dialog checks that the key resolves and warns you if it cannot be found.
The matching Credential must be tenant-scoped (available to all organizations) for the engine to resolve it. A credential restricted to a specific organization will not be found. In the Credentials dialog, leave the Organization field unset so the credential is created as Tenant, not Org.
Providers that do not need a key, for example a self-hosted model on a private network, can turn off Requires auth.

​
Azure settings

For the azure type, two extra fields appear:
FieldDefaultDescription
API Version2025-04-01-previewThe Azure OpenAI API version to target
API ModechatWhich API surface to call: chat, responses, or completion
Azure API Management (APIM) gateways are supported as the endpoint: set the Base URL to your APIM gateway URL to route all model traffic through it for centralized policy, rate limiting, and logging.
A provider that still has models cannot be deleted. Remove or reassign its models first, or simply disable the provider instead.

​
Models

Each model belongs to a provider. Open Settings > Models and click Add Model.
FieldRequiredDescription
ProviderYesThe provider that serves this model
Model IDYesMust match the model’s deployment name on the provider exactly—this is the identifier sent to the provider’s API, not a label you choose
KeyNoThe display name shown in Command Center and referenced by agents when selecting an engine. Auto-derived from the Model ID if left blank. To rename a model, edit this field directly
Context lengthYesMaximum tokens the model accepts. Drives the context-management guidance in Engines
Accepts image inputNoEnable for models that can read images. Image-capable models become selectable as the tenant’s Multimodal model (see Tenant defaults)
EnabledNoOnly enabled models appear as selectable engines. Disable a model to retire it without losing its configuration or history
Advanced sampling parameters (temperature, top P, top K, presence/frequency penalty, max output tokens, seed, stop sequences) can be set per model. You can also record pricing plans per model and run an on-demand performance test that reports reachability, time to first token, and throughput.

​
Thinking effort

A thinking-effort level trades depth of reasoning against latency and cost:
LevelDescription
OffNo reasoning tokens—fastest, lowest cost
MinimalVery light reasoning
LowBasic reasoning pass
MediumBalanced reasoning depth
HighDeep reasoning for complex tasks
xHighMaximum reasoning; highest cost
Some custom or compatible providers do not accept a reasoning argument at all and will reject any request that includes one. If your provider does not support reasoning, leave thinking effort unset—do not select any level.
A model assigned to any agent—as a main or secondary engine—cannot be deleted. Reassign those agents to other engines, or disable the model instead.

​
Tenant defaults

The tenant’s default Text LLM model and Multimodal model are chosen in Settings > Tenant from the catalog’s enabled models. Only image-capable models appear as Multimodal options. These defaults power agents that do not specify their own. See Engines for how agents select the models that power them.

​
Deploy-time seeding

On deploy, the installer seeds providers and models from the API keys detected in the engine environment. After that, management moves to Command Center.