Reference
AI Models
The AI models the server can use, such as for spam filtering and “Explain this”.
On this page
In the console: Settings › System › AI models
The AI models the server can use, such as for spam filtering and “Explain this”.
AI Endpoint settings
Name
Short name for the AI Model.
Text. Name in the API: name.
Endpoint URL
URL of the OpenAI compatible endpoint.
An address (URL). Name in the API: url.
Allow Invalid Certs
Whether the server should connect to an endpoint that has an invalid TLS certificate.
On or off. Name in the API: allowInvalidCerts.
Model
Type
API type.
One of: Chat Completion (Chat), Text Generation (Text). Default: Chat Completion. Name in the API: modelType.
Model
The name of the AI model to use.
Text. Name in the API: model.
Temperature
The temperature of the AI model, which controls the randomness of the output. A higher temperature will produce more random output.
A number. Default: 0.7. Name in the API: temperature.
Authentication
Authentication
The type of HTTP authentication to use.
Settings of their own: see HTTP auth. Name in the API: httpAuth.
Options
Timeout
Maximum amount of time that the server will wait for a response from this endpoint.
A length of time. Default: 2m. Name in the API: timeout.
HTTP Headers
Additional headers to include in HTTP requests.
A set of entries, each a key and its value. Name in the API: httpHeaders.