Skip to main content
GET
List available models
Returns OpenAI-compatible payload:

Filter by modality

There is no modality query parameter — the endpoint takes only include_tools, and any other query key is ignored rather than refused. List once and filter client-side on _infery.modality:
Modality values: text, image, audio, stt, video, music, embedding, upscale, vision.

Capabilities

There is no per-model GET /v1/models/{slug} — the list response already carries everything, under each entry’s _infery object:
  • supports_chat / supports_streaming / supports_tools / supports_embeddings
  • max_context_tokens / max_output_tokens
  • pricing — retail rates in credits
  • allowed_params — the generation knobs the model accepts: sizes, image_sizes, aspect_ratios, resolutions, qualities, voices, durations, response_formats, max_n, default_params
Fetch the list once at app startup and index it by id to learn what each model can do. To price a specific call before making it, POST /v1/models/{slug}/estimate.

Public catalog

No-auth version available at:
Rate-limited (30 rpm per IP). Used by our marketing site to show a live model list.

Authorizations

Authorization
string
header
required

API key in format: Bearer inf_***

Query Parameters

include_tools
string

Include kind='tool' rows (default: false)

Response

List of available models. OpenAI-compatible with an _infery extension exposing capabilities and pricing.

object
string
Example:

"list"

data
object[]