Skip to main content
POST
Create a new AI model

Authorizations

x-access-token
string
header
required

JWT token passed in x-access-token header

Body

application/json
name
string
required

Model name (must be unique per institution + model_use).

institution
string

Institution ID. Defaults to the caller's institution when omitted.

model_use
enum<string>
default:conversationModel
Available options:
conversationModel,
imageGenerationModel,
audioGenerationModel,
videoGenerationModel,
imageAnalysisModel,
embeddingsModel,
audioAnalysisModel,
videoAnalysisModel,
ttsModel,
summaryModel,
rtModel,
moderationModel
description
string
client_library
enum<string>
default:openai_cli

SDK / dispatch backend used to talk to the provider.

Available options:
openai_cli,
anthropic_cli,
bedrock_cli,
google_cli,
mistral_cli,
xai_cli,
stability_cli
api_url
string

Provider base URL. Required for openai_cli custom models.

api_key
string

Provider API key. Required for openai_cli, mistral_cli, and xai_cli custom models.

default_headers
object
features
object

Capability flags advertised by the model.

input_media
object

Input modalities the model accepts.

output_media
object

Output modalities the model can produce.

max_input_tokens
integer
default:128000
max_output_tokens
integer
default:64000
reasoning_effort
enum<string> | null
Available options:
none,
low,
medium,
high,
max
extended_context
boolean
default:false
status
enum<string>
default:active
Available options:
active,
inactive
picture_url
string

Response

AI model created successfully

success
boolean
id
string

The new AI model's ObjectId.

message
string
Example:

"AIModel created!"