UapiProUApiPro
AI & Models

Query the AI model catalog

0calls
2 credits / call

Want to build a dynamic "model picker" into your AI application? Or compare the parameters and billing prices of LLMs across aggregation platforms? This endpoint gives you the full picture of mainstream AI models in one call.

GET
uapis.cn
/api/v1/ai/models
Query parameters
9
q
string

Global fuzzy search keyword, matching model ID, name, provider, or description.

provider
string

Filter by the model's provider. Supports comma-separated multi-select, e.g. deepseek,openai.

type
string

Filter by the model's base use case type.

modality
string

Filter by supported modality. Matches if either the input or the output supports the modality.

capability
string

Filter by advanced capability.

sort
string

Sort field. Uses the default order when omitted.

order
string

Sort direction: ascending (asc) or descending (desc).

page
integer

The page number to request. Defaults to 1.

limit
integer

Maximum number of models to return. Range 0 - 5000. Defaults to 0, meaning no pagination cap and all matching models are returned.

Overview

Calling this endpoint returns a model catalog aggregated from multiple sources (such as OpenRouter, Vercel, and more). Along with each model's ID, name, and provider, it details the supported input/output modalities (text, image, file, etc.) and advanced capabilities such as deep reasoning and tool calling.

Multi-dimensional filtering and sorting Use the query parameters to precisely trim the large catalog. For example, filter models from the deepseek provider, or only keep models that support the image modality. You can also sort by release date or context window size.

Pagination and size control Set limit to 0 (the default) to pull the entire catalog. For display or testing purposes, we recommend passing a reasonable limit value.

Query parameters

q
stringoptional

Global fuzzy search keyword, matching model ID, name, provider, or description.

provider
stringoptional

Filter by the model's provider. Supports comma-separated multi-select, e.g. deepseek,openai.

type
stringoptional

Filter by the model's base use case type.

modality
stringoptional

Filter by supported modality. Matches if either the input or the output supports the modality.

capability
stringoptional

Filter by advanced capability.

sort
stringoptional

Sort field. Uses the default order when omitted.

order
stringoptional

Sort direction: ascending (asc) or descending (desc).

page
integeroptional

The page number to request. Defaults to 1.

limit
integeroptional

Maximum number of models to return. Range 0 - 5000. Defaults to 0, meaning no pagination cap and all matching models are returned.

Response

200 / OK

Query succeeded. Returns the matching model list along with pagination metadata.

JSON
{
  // List of models.
  "data": [
    {
      // Unique model identifier.
      "id": "deepseek-v4-pro",
      // Display name of the model.
      "name": "DeepSeek V4 Pro",
      // Provider name.
      "provider": "deepseek",
      // Model type.
      "type": "language",
      // Supported input modalities.
      "input_modalities": [
        "string"
      ],
      // Supported output modalities.
      "output_modalities": [
        "string"
      ],
      // Advanced capability tags of the model.
      "capabilities": [
        "string"
      ]
    }
  ],
  // Pagination and status metadata.
  "meta": {
    // Total number of matching models.
    "total": 13,
    // Number of models returned in this response.
    "returned": 13,
    // Current page number.
    "page": 1,
    // Maximum items per request.
    "limit": 100,
    // Whether another page exists.
    "has_more": false,
    // Timestamp of the catalog snapshot.
    "generated_at": "2026-08-06T14:42:30Z",
    // Whether the current catalog data is stale (not updated for a long time).
    "stale": false
  }
}

400 / Bad Request

Invalid query parameters, e.g. limit out of range, or an unsupported modality/sort type.

JSON
{
  "code": "INVALID_PARAMETER",
  "message": "The specified parameter is invalid"
}

502 / Bad Gateway

The model catalog aggregation service is temporarily unavailable. Please retry later.

JSON
{
  "code": "MODEL_CATALOG_UNAVAILABLE",
  "message": "The model catalog service is temporarily unavailable, please retry later"
}

504 / Gateway Timeout

The model catalog timed out while processing a large amount of model data. Please retry later.

JSON
{
  "code": "MODEL_CATALOG_TIMEOUT",
  "message": "The model catalog timed out, please retry later"
}
PreviousCategory overview