Skip to content

Which models are actually in use, and whether any is hosted

GET
/api/v1/models/in-use
curl --request GET \
--url http://localhost:11044/api/v1/models/in-use \
--header 'X-Embabel-Api-Key: <X-Embabel-Api-Key>'

Models in use, with providers

Media typeapplication/json

Which models the caller’s chat, documents and embeddings actually use, and whether any is hosted; answered by GET /models/in-use

object
allLocal
required

True when every surface that resolved a model runs it on this machine; surfaces with no model do not count against it

boolean
cloudProviders
required

Distinct names of the hosted (non-local) providers in use, sorted; empty when allLocal is true

Array<string>
models
required

One entry per surface, in the order chat, documents, embeddings

Array<object>

One surface of the appliance (chat, documents, embeddings) and the model that serves it, as GET /models/in-use resolves it

object
local
required

True when the provider is one that runs on this machine (LM Studio, Ollama, Docker); false otherwise, including when the provider is unknown

boolean
model

Model name serving this surface, resolved the way the surface itself resolves it; null when none could be resolved

string
provider

Provider of the model, as the service serving it names it (e.g. “LM Studio”, “OpenAI”); null when there is no model

string
purpose
required

What the model is used for, in the user’s terms, as “ — ”: currently “Chat — what you talk to”, “Documents — answering from your files” or “Embeddings — indexing and search”

string
summary
required

One plain-English sentence saying whether anything goes to a model provider, and which surface uses which hosted model

string

Example generated

{
"allLocal": true,
"cloudProviders": [
"example"
],
"models": [
{
"local": true,
"model": "example",
"provider": "example",
"purpose": "example"
}
],
"summary": "example"
}