Create AI Model

Used to create a new AI model with specified configuration. This allows you to integrate various AI models from different providers into your project.

API Endpoint

Property	Value
Request Method	POST
Request URL	https://api.seliseblocks.com/models/

Request

Request Example

curl -X POST 'https://api.seliseblocks.com/models/' \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '{
    "provider": "openai",
    "service_platform": "official_api",
    "version": "gpt-4",
    "model_name": "gpt-4-turbo-preview",
    "project_key": "YOUR_PROJECT_KEY",
    "display_name": "GPT-4 Turbo",
    "api_key": "sk-your-api-key-here",
    "base_url": "https://api.openai.com/v1",
    "MaxTokens": 8192,
    "DefaultTemp": 0.7
  }'

Request Headers

Field	Type	Required	Description
accept	string	Yes	Accepted response format. Use application/json
Content-Type	application/json	Yes	Data type, must be application/json.

Request Body

Request Body Schema

{
  "provider": "openai",
  "service_platform": "official_api",
  "version": "string",
  "model_name": "string",
  "project_key": "string",
  "display_name": "string",
  "api_key": "string",
  "base_url": "string",
  "openai_organization_id": "string",
  "openai_project_id": "string",
  "api_version": "string",
  "deployment_name": "string",
  "custom_parameters": {
    "additionalProp1": {}
  },
  "custom_headers": {
    "additionalProp1": {}
  },
  "MaxTokens": 8192,
  "DefaultTemp": 0.7
}

Request Body Parameters

Field	Type	Required	Description
provider	string	Yes	AI model provider (e.g., "openai", "anthropic", "azure", "google").
service_platform	string	Yes	Service platform type (e.g., "official_api", "azure", "custom").
version	string	Yes	Model version identifier.
model_name	string	Yes	Specific model name (e.g., "gpt-4-turbo-preview", "claude-3-opus").
project_key	string	Yes	The project key for your project.
display_name	string	Yes	Human-readable display name for the model.
api_key	string	Yes	API key for authentication with the provider.
base_url	string	No	Base URL for API requests (defaults to provider's standard URL).
openai_organization_id	string	No	OpenAI organization ID (for OpenAI models only).
openai_project_id	string	No	OpenAI project ID (for OpenAI models only).
api_version	string	No	API version (required for Azure OpenAI).
deployment_name	string	No	Deployment name (required for Azure OpenAI).
custom_parameters	object	No	Additional custom parameters for the model.
custom_headers	object	No	Custom HTTP headers for API requests.
MaxTokens	integer	No	Maximum number of tokens the model can generate (default: 8192).
DefaultTemp	number	No	Default temperature for generation (default: 0.7, range: 0-2).

note

Provider-Specific Requirements

• OpenAI: Requires api_key, optionally openai_organization_id and openai_project_id
• Azure OpenAI: Requires api_key, base_url, api_version, and deployment_name
• Anthropic: Requires api_key, model name like "claude-3-opus-20240229"
• Google: Requires appropriate authentication credentials in api_key
• Custom providers: May require base_url and custom_headers

tip

Best practices for model configuration:

• Use descriptive display names to easily identify models
• Store API keys securely and rotate them regularly
• Set appropriate MaxTokens based on your use case and budget
• Start with DefaultTemp of 0.7 and adjust based on output quality
• Test model configuration after creation
• Document model-specific parameters in custom_parameters
• Use separate models for development and production

Temperature guidelines:

• 0.0-0.3: More deterministic, factual responses
• 0.4-0.7: Balanced creativity and consistency (recommended default)
• 0.8-1.0: More creative and varied responses
• 1.1-2.0: Highly creative, potentially less coherent

The system will:

• Validate API credentials during creation
• Test connectivity to the model provider
• Store credentials securely (encrypted at rest)
• Make the model available for AI agents immediately
• Track model usage and costs

Response

Success Response (201 Created)

Returns an object containing the creation status.

{
  "is_success": true,
  "item_id": "model_123",
  "detail": "AI model created successfully",
  "error": {}
}

Response Fields

Field	Type	Description
is_success	boolean	Indicates whether the creation was successful.
item_id	string	The unique identifier of the created model.
detail	string	Success or failure message with additional context.
error	object	Error details if the operation failed (empty if successful).

Error Response (422 Unprocessable Entity)

Returns validation error details when the request body is invalid.

{
  "detail": [
    {
      "loc": [
        "body",
        "api_key"
      ],
      "msg": "Invalid API key format or authentication failed",
      "type": "value_error.authentication"
    }
  ]
}

Error Response Fields

Field	Type	Description
detail	array	Array of validation error objects.
loc	array	Location of the error in the request (e.g., body field).
msg	string	Human-readable error message.
type	string	Error type identifier.

Common Error Scenarios

Error Type	Description	Resolution
Invalid API Key	API key is incorrect or expired	Verify API key from provider dashboard
Authentication Failed	Provider rejected credentials	Check API key permissions and quotas
Invalid Model Name	Model name doesn't exist or is misspelled	Verify model name with provider documentation
Missing Required Field	Required provider-specific field missing	Add required fields (e.g., deployment_name for Azure)
Connection Failed	Cannot connect to provider's API	Check base_url and network connectivity
Quota Exceeded	Provider API quota exceeded	Upgrade plan or wait for quota reset

Error Codes

Status Code	Description	Response Type
201	Created - Model created successfully	Success
400	Bad Request - Invalid model configuration	Bad Request
401	Unauthorized - Invalid API credentials	Unauthorized
409	Conflict - Model with same name already exists	Conflict
422	Validation Error - Invalid request parameters	Unprocessable Entity
503	Service Unavailable - Provider API unavailable	Service Unavailable