LLM API

This API provides endpoints to retrieve information about available LLMs. It allows you to obtain a list of providers, their models, and model-specific details.

Check the Glob.AI OS API Reference for generic variables needed to use the API.

Endpoints

Authentication

All endpoints require authentication using one of the following:

• Authorization: Bearer $OAuth_accesstoken
• Authorization: Bearer $GEAI_ORGANIZATION_APITOKEN
• Authorization: Bearer $GEAI_APITOKEN

For $OAuth_accesstoken, you must also include the header: project-id: $GEAI_PROJECT_ID. If a $GEAI_APITOKEN is provided, the header project-id is not required.

Additionally, all endpoints can include this optional header:

• Accept-Language: desired language for Provider and Model description. (For example: English, Spanish, Japanese). If it is not declared, English will be set as the default language.

GET /llm/providers

Retrieves a list of available LLM providers. Optionally includes model details.

Parameters

Note: The model-related filters (type, streaming, multimodal, toolCalling, functionCalling) apply only when include=models.

Request

• Method: GET
• Path: $BASE_URL/v3/llm/providers
• Body: Empty

Response

A successful response (200 OK) returns a JSON object with an array of providers. Each provider object contains at least a description, interfaceType, and name. If include=models is specified, each provider will also include a models array.  Each model object contains details like contextWindow, fullName, id, isCustom, maxOutputTokens, name, priority, properties, and type.

{
  "providers": [                                  // array of provider objects
    {
      "description": "string",                    // Provider description (localized if requested)
      "friendlyName": "string",                   // Human-friendly provider name
      "interfaceType": "string",                  // Interface style (e.g., "OpenAI")
      "name": "string"                            // Provider internal name (e.g., "Anthropic")
    },
    {
      "description": "string",
      "friendlyName": "string",
      "interfaceType": "string",
      "name": "string"
    }
  ]
}

{
  "providers": [                                      // array of all providers
    {
      "name": "string",                               // System identifier for the provider (e.g., "openai", "azure")
      "friendlyName": "string",                       // Human-friendly name (e.g., "OpenAI", "Azure OpenAI")
      "description": "string",                        // Summary of the provider's positioning/capabilities
      "interfaceType": "string",                      // Underlying API interface (e.g., "openai")
      "models": [                                     // List of models available from this provider
        {
          "id": "string",                             // Model GUID (internal unique identifier)
          "name": "string",                           // Short model name (e.g., "gpt-4o")
          "fullName": "string",                       // Fully qualified model name (e.g., "openai/gpt-4o")
          "description": "string",                    // Detailed model description and release notes
          "type": "string",                           // Model interface type (e.g., "Chat")
          "contextWindow": integer,                   // Maximum input tokens supported by the model
          "maxOutputTokens": integer,                 // Maximum output tokens supported by the model
          "functionCalling": boolean,                 // Supports function calling (structured tool calls)
          "toolCalling": boolean,                     // Supports tool calling
          "streaming": boolean,                       // Supports server-sent streaming responses
          "multimodal": boolean,                      // Supports multimodal inputs (e.g., images, files, audio, video)
          "isCustom": boolean,                        // Indicates if this is a custom-deployed model
          "priority": integer,                        // Optional model ranking or routing priority
          "parameters": {},                           // Provider/model-specific tunables (e.g., temperature policies)
          "mediaCapabilities": {                      // Per-modality limits and MIME types
            "file": {
              "maxFileSizeMB": integer,               // Maximum file size in MB
              "maxFilesPerPrompt": integer,           // Max number of files allowed per request (0 = unlimited or not enforced)
              "maxPagesPerFile": integer,             // Max pages per file (0 = unlimited or not enforced)
              "mimeTypes": [ "string" ]               // Allowed file MIME types
            },
            "image": {
              "maxImageSizeMB": integer,              // Maximum image size in MB
              "maxImagesPerPrompt": integer,          // Max images allowed per request
              "mimeTypes": [ "string" ]               // Allowed image MIME types
            },
            "audio": {
              "maxAudioLengthMinutes": integer,       // Maximum audio length in minutes
              "maxFilesPerPrompt": integer,           // Max audio files per request
              "mimeTypes": [ "string" ]               // Allowed audio MIME types
            },
            "video": {
              "maxVideosPerPrompt": integer,          // Max videos per request
              "maxLengthWithAudioMinutes": integer,   // Max video length with audio (minutes)
              "maxLengthWithoutAudioMinutes": integer,// Max video length without audio (minutes)
              "mimeTypes": [ "string" ]               // Allowed video MIME types
            }
          },
          "properties": [                             
            {
              "id": "string",                         // Property GUID
              "name": "string",                       // Capability or attribute name (e.g., "SupportsToolChoice")
              "type": "string",                       // Property type label (e.g., "Boolean", "String")
              "value": "string"                       // Property value serialized as string (e.g., "true", CSV list)
            }
          ]
        }
      ]
    }
  ]
}

Error Responses

• 401 Unauthorized: Invalid OAuth access token or API token. Response body: {"errors":[{"description":"Unauthorized","id":401}]}

• 404 Not Found: No data available. Response body: {"errors":[{"description":"Not Found.","id":4000}]}

cURL Sample

curl -X GET '$BASE_URL/v3/llm/providers' \
-H 'Content-Type: application/json' \
-H 'Authorization: Bearer $GEAI_APITOKEN'

curl -X GET '$BASE_URL/v3/llm/providers?include=models&type=Chat&functionCalling=true&toolCalling=true&multimodal=true&streaming=true' \
-H 'Content-Type: application/json' \
-H 'Accept-Language: english' \
-H 'Authorization: Bearer $GEAI_APITOKEN'

GET /llm/providers/{providerName}

Retrieves specific data for a given LLM provider.

Parameters

Request

• Method: GET
• Path: $BASE_URL/v3/llm/providers/{providerName}
• Body: Empty

Response  

If the endpoint execution is successful, it will return a Status 200 OK with the requested data.

{
  "description": "string",                        // Provider description
  "friendlyName": "string",                       // Human-friendly provider name
  "interfaceType": "string",                      // Interface style (e.g., "openai")
  "name": "string"                                // Provider internal name (e.g., "openai")
}

Error responses

• 401 Unauthorized: Invalid OAuth access token or API token. Response body: {"errors":[{"description":"Unauthorized","id":401}]}
• 404 Not Found: No data available. Response body: {"errors":[{"description":"Not Found.","id":4000}]}.

cURL Sample

curl -X GET '$BASE_URL/v3/llm/providers/openai' \
-H 'Content-Type: application/json' \
-H 'Authorization: Bearer $GEAI_APITOKEN'

GET /llm/providers/{providerName}/models

Retrieves a list of models for a specific provider.

Parameters

Request

• Method: GET
• Path: $BASE_URL/v3/llm/providers/{providerName}/models
• Body: Empty

Response

If the endpoint execution is successful, it will return a Status 200 OK with the requested data.

{
  "models": [
    {
      "id": "string",                               // Globant model ID (GUID)
      "name": "string",                             // Short model name (e.g., "gpt-4o")
      "fullName": "string",                         // Provider-qualified name (e.g., "openai/gpt-4o")
      "description": "string",                      // Human-readable model description
      "type": "string",                             // Model interface type (e.g., "Chat")
      "priority": integer,                          // Ordering hint for UI/catalogs
      "isCustom": boolean,                          // True if this is a custom/finetuned model
      "multimodal": boolean,                        // True if model supports multiple modalities
      "streaming": boolean,                         // True if server-sent streaming is supported
      "toolCalling": boolean,                       // True if model supports tool/function calling
      "functionCalling": boolean,                   // True if function-calling API is supported
      "contextWindow": integer,                     // Max input tokens the model can accept
      "maxOutputTokens": integer,                   // Max tokens the model can generate in one response

      "mediaCapabilities": {                        // Per-modality constraints
        "file": {
          "maxFileSizeMB": integer,                 // Max file size in MB
          "maxFilesPerPrompt": integer,             // Max number of files per prompt (0 = unlimited/not enforced)
          "maxPagesPerFile": integer,               // Max pages per file (for paged docs like PDFs)
          "mimeTypes": [ "string" ]                 // Supported file MIME types
        },
        "image": {
          "maxImageSizeMB": integer,                // Max image size in MB
          "maxImagesPerPrompt": integer,            // Max images per prompt (0 = unlimited/not enforced)
          "mimeTypes": [ "string" ]                 // Supported image MIME types
        }
      },

      "parameters": {                               // Model-tunable parameters (if any)
        "temperature": {
          "supported": boolean,                     // Whether parameter is supported
          "allowedValues": [ number ],              // Enumerated allowed values (if constrained)
          "range": [ number ]                       // Continuous range (if applicable) - may be empty
        },
        "top_p": {
          "supported": boolean                      // Whether parameter is supported
        }
        /* Additional parameter keys may appear depending on the model */
      },

      "properties": [                               // Additional provider- or system-defined properties
        {
          "id": "string",                           // Property ID (GUID)
          "name": "string",                         // Property name (e.g., "SupportsMultimodal")
          "type": "string",                         // Data type of value, e.g., "Boolean" | "String"
          "value": "string"                         // Property value encoded as string
        }
      ]
    }
  ]
}

Error responses

• 401 Unauthorized: If the OAuth access token or API token are invalid. Response body: {"errors":[{"description":"Unauthorized","id":401}]}.
• 404 Not Found: If there is no data available. Response body: {"errors":[{"description":"Not Found.","id":4000}]}.

cURL Sample

curl -X GET '$BASE_URL/v3/llm/providers/openai/models' \
-H 'Content-Type: application/json' \
-H 'Accept-Language: english' \
-H 'Authorization: Bearer $GEAI_APITOKEN'

GET /llm/providers/{providerName}/models/{modelNameOrId}

Retrieves data for a specific model.

Parameters

Request

• Method: GET
• Path: $BASE_URL/v3/llm/providers/{providerName}/models/{modelNameOrId}
• Body: Empty

Response

If the endpoint execution is successful, it will return a Status 200 OK with the requested data.

{
  "description": "string"     // Provider description and intended use
  "friendlyName": "string",   // Human-friendly display name
  "interfaceType": "string",  // Connector/interface identifier (e.g., "openai")
  "name": "string",           // Unique provider key (e.g., "openai")
}

Error Responses

• 401 Unauthorized: Invalid OAuth access token or API token. Response body: {"errors":[{"description":"Unauthorized","id":401}]}
• 404 Not Found: No data available. Response body: {"errors":[{"description":"Not Found.","id":4000}]}
• 409 Conflict: Multiple models share the same name. Response body: {"errors":[{"description":"The model name is ambiguous.","id":4000}]} Recommendation: Use the id instead of the name.

cURL Sample

curl -X GET '$BASE_URL/v3/llm/providers/openai/models/gpt-4o' \
-H 'Content-Type: application/json' \
-H 'Accept-Language: english' \
-H 'Authorization: Bearer $GEAI_APITOKEN'