add-ai-model
Adds a new AI model configuration.
add-ai-model
[-b value | --bootstrap-config=value]
[-t value | --tool-password=value]
<--name=value>
<--model=value>
<--api=value>
[--mode=value]
<--api-base=value>
<--auth-type=value>
[--api-key=value]
[--priority=value]
[--disable]
{-Gvalue}
{-Uvalue}
[--auth-header-mode=value]
[--auth-header=value]
[--path-override=value]
{-Hkey=value}
<-u value | --user=value>
Overview
Use this command to modify an existing AI model configuration. Only the arguments that are explicitly provided are changed; omitted arguments leave the corresponding fields unchanged.
Options
| Option | Optional or Required | Default Value | Description |
|---|---|---|---|
|
Optional | none | The path to the bootstrap configuration file. See Bootstrap.xml file for more information about this file. |
|
Optional | none | The configuration tool password used to decrypt the database password in the file bootstrap.xml. If the tool password is omitted, the command prompts the user for it in the console. See Bootstrap.xml file. |
|
Required | none | The display name of the AI model. Must be unique within the given scope. |
|
Required | none | The model identifier string. For example,
gpt-4o or
claude-3-5-sonnet.
|
|
Required | none | The API type for the model. The valid
values are
|
|
Optional | Completion
|
The model interaction mode. The valid value
is
Completion.
|
|
Required | none | The base URL of the AI provider endpoint.
For example,
https://api.openai.com/v1
|
|
Required | none | The authentication type used to call the AI
provider. The valid values are
ApiKey and
None. Use
None for providers that do not require
authentication, for example local Ollama instances.
|
|
Optional | none | The API key used for autentication.
Ifauth-type=ApiKey and this setting is omitted, the command
will prompt for a key. Must not be specified when
auth-type=None.
|
|
Optional | none |
The priority of this model relative to others in this scope. Lower values indicate higher priority. |
|
Optional | none | If specified, the model is created in a disabled state. Disabled models are excluded from user-facing listings and proxying. |
|
Optional | none | The name of a group whose members are allowed to use this model. Can be specified multiple times with different values to allow multiple groups. |
|
Optional | none | The name of a user who is individually allowed to use this model. Can be specified multiple times with different values to allow multiple users. |
|
Optional | Default
|
Controls how the authorization header is
constructed. The valid values are
Default (use the convention defined by the API
type),
AuthorizationBearer (always send
Authorization: Bearer <apiKey>) and
Custom (send
<auth-header>: <apiKey>).
|
|
Optional | none | The custom authorization header name to use
when
--auth-header-mode=Custom. The header value
will be the raw API key. Required if
--auth-header-mode=Custom.
|
|
Optional | none |
An optional path override for clients. When set, clients use this path instead of the default when constructing gateway requests. |
|
Optional | none |
A custom HTTP header to set on the model, specified as 'name=value'. Can be specified multiple times with different keys to add multiple HTTP headers. |
|
Required | none | The user performing the operation. This should be an administrator. Unless the user is part of the configured default domain, the name of the user must include the user's domain name. For example, DOMAIN\user or user@domain. |