Models
The model catalogue lists the language models your deployment offers, as configured by your administrator. Use a model's model_name when you choose the model for an agent.
Each model belongs to a provider, the label the catalogue groups it under. Filter List models by provider to see one group, and retrieve a model for its context window, costs and supported features.
Paths are relative to /agent-hub-api
The model object
One catalogue entry.
Attributes
-
model_namestringUnique identifier of the model in the catalogue. May contain
/. -
namestringDisplay name of the model.
-
providerstringLabel the catalogue groups the model under.
-
typestringModel type from the catalogue, for example
chat. -
context_windownullable integerMaximum context length in tokens, when known.
-
max_tokensnullable integerMaximum number of output tokens, when known.
-
cost_per_1k_inputnullable numberCost of 1,000 input tokens, when known.
-
cost_per_1k_outputnullable numberCost of 1,000 output tokens, when known.
-
supports_toolsbooleanWhether the model can call tools.
-
supports_streamingbooleanWhether the model can stream its output.
-
supports_visionbooleanWhether the model accepts images.
-
supports_json_modebooleanWhether the model supports structured JSON output.
-
supports_reasoningbooleanWhether the model is a reasoning model.
-
temperature_supportedbooleanWhether the model accepts a
temperaturesetting. -
capabilitiesarray of stringsCapability labels, sorted, such as
reasoning,tool_useandvision. -
tagsarray of stringsFree-form tags from the catalogue.
-
levelnullable stringLevel label from the catalogue, when set.
-
regulated_approvedbooleanWhether the model is approved for use in regulated domains.
{
"provider": "meta-llama",
"model_name": "meta-llama/llama-3.3-70b-instruct",
"name": "Llama 3.3 70B Instruct",
"type": "chat",
"context_window": 131072,
"max_tokens": 16384,
"cost_per_1k_input": 0.0001,
"cost_per_1k_output": 0.00032,
"supports_tools": true,
"supports_streaming": true,
"supports_vision": false,
"supports_json_mode": true,
"supports_reasoning": false,
"temperature_supported": true,
"capabilities": [
"tool_use"
],
"tags": [
"open-weights"
],
"level": null,
"regulated_approved": true
} List models
GET /agent-hub-api/api/models
Returns the models in the catalogue, sorted by model_name.
Returns plain model names by default. Set detailed to get a short summary of each model; retrieve a model for all of its attributes.
- Authentication
- Bearer token How it works
Query parameters
-
providerstringOnly models whose
provideris exactly this value. An unknown provider returns an empty list. -
detailedbooleanSet to
trueto return objects withmodel_name,name,providerandtypeinstead of plain names.
Returns
Returns data, a list of model names, or of model summaries when detailed is true.
Errors
- 503 The model catalogue could not be loaded or is empty.
curl "$VDF_BASE_URL/agent-hub-api/api/models?detailed=true" \
-H "Authorization: Bearer $VDF_ACCESS_TOKEN" const response = await fetch(`${process.env.VDF_BASE_URL}/agent-hub-api/api/models?detailed=true`, {
headers: {
Authorization: `Bearer ${process.env.VDF_ACCESS_TOKEN}`,
},
});
if (!response.ok) throw new Error(`Request failed with status ${response.status}`);
const data = await response.json(); import os
import requests
response = requests.get(
f"{os.environ['VDF_BASE_URL']}/agent-hub-api/api/models",
headers={"Authorization": f"Bearer {os.environ['VDF_ACCESS_TOKEN']}"},
params={
"detailed": "true",
},
timeout=30,
)
response.raise_for_status()
data = response.json() {
"success": true,
"data": [
{
"model_name": "meta-llama/llama-3.3-70b-instruct",
"name": "Llama 3.3 70B Instruct",
"provider": "meta-llama",
"type": "chat"
},
{
"model_name": "mistralai/mistral-small-3.2-24b-instruct",
"name": "Mistral Small 3.2 24B Instruct",
"provider": "mistralai",
"type": "chat"
}
]
} Retrieve a model
GET /agent-hub-api/api/models/{name}
Returns one model from the catalogue.
- Authentication
- Bearer token How it works
Path parameters
-
namestring RequiredThe model's
model_name. It may contain/, which you can send as is.
Returns
Returns data, the model object.
Errors
- 404 No model with this name is in the catalogue.
- 503 The model catalogue could not be loaded or is empty.
curl "$VDF_BASE_URL/agent-hub-api/api/models/meta-llama/llama-3.3-70b-instruct" \
-H "Authorization: Bearer $VDF_ACCESS_TOKEN" const response = await fetch(`${process.env.VDF_BASE_URL}/agent-hub-api/api/models/meta-llama/llama-3.3-70b-instruct`, {
headers: {
Authorization: `Bearer ${process.env.VDF_ACCESS_TOKEN}`,
},
});
if (!response.ok) throw new Error(`Request failed with status ${response.status}`);
const data = await response.json(); import os
import requests
response = requests.get(
f"{os.environ['VDF_BASE_URL']}/agent-hub-api/api/models/meta-llama/llama-3.3-70b-instruct",
headers={"Authorization": f"Bearer {os.environ['VDF_ACCESS_TOKEN']}"},
timeout=30,
)
response.raise_for_status()
data = response.json() {
"success": true,
"data": {
"provider": "meta-llama",
"model_name": "meta-llama/llama-3.3-70b-instruct",
"name": "Llama 3.3 70B Instruct",
"type": "chat",
"context_window": 131072,
"max_tokens": 16384,
"cost_per_1k_input": 0.0001,
"cost_per_1k_output": 0.00032,
"supports_tools": true,
"supports_streaming": true,
"supports_vision": false,
"supports_json_mode": true,
"supports_reasoning": false,
"temperature_supported": true,
"capabilities": [
"tool_use"
],
"tags": [
"open-weights"
],
"level": null,
"regulated_approved": true
}
}