Upsert Seed Models for Provider

Used to create or update the seed models data for a specific AI provider. This endpoint allows you to add new models or update existing model specifications in the seed collection.

API Endpoint

Property	Value
Request Method	POST
Request URL	https://api.seliseblocks.com/models/seed/providers/{provider}/models

Request

Request Example

curl -X POST 'https://api.seliseblocks.com/models/seed/providers/openai/models' \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '[
    {
      "Model": "gpt-4-turbo-preview",
      "ModelGoodName": "GPT-4 Turbo",
      "MaxTokens": 4096,
      "DefaultTemp": 0.7,
      "ContextLength": 128000,
      "InputCostPerMillion": 10.0,
      "OutputCostPerMillion": 30.0
    },
    {
      "Model": "gpt-3.5-turbo",
      "ModelGoodName": "GPT-3.5 Turbo",
      "MaxTokens": 4096,
      "DefaultTemp": 0.7,
      "ContextLength": 16385,
      "InputCostPerMillion": 0.5,
      "OutputCostPerMillion": 1.5
    }
  ]'

Path Parameters

Field	Type	Required	Description
provider	string	Yes	Provider name (case-insensitive). Examples: "openai", "anthropic", "azure", "google".

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

[
  {
    "Model": "string",
    "ModelGoodName": "string",
    "MaxTokens": 0,
    "DefaultTemp": 0,
    "ContextLength": 0,
    "InputCostPerMillion": 0,
    "OutputCostPerMillion": 0
  }
]

Request Body Parameters

Field	Type	Required	Description
Model	string	Yes	Exact model identifier used in API calls (e.g., "gpt-4-turbo-preview").
ModelGoodName	string	Yes	Human-readable display name for the model (e.g., "GPT-4 Turbo").
MaxTokens	integer	Yes	Maximum number of tokens the model can generate in a single response.
DefaultTemp	number	Yes	Default temperature setting for the model (typically 0-2).
ContextLength	integer	Yes	Maximum context window size in tokens (input + output).
InputCostPerMillion	number	Yes	Cost per million input tokens in USD.
OutputCostPerMillion	number	Yes	Cost per million output tokens in USD.

note

Upsert Behavior

This endpoint performs an upsert operation:

• Create: If a model with the given Model identifier doesn't exist, it will be created
• Update: If a model with the given Model identifier already exists, it will be updated with the new values
• Replace: The entire models array for the provider is replaced, so include all models you want to keep

Provider-Specific Considerations:

• Ensure model names match the exact identifiers used by the provider's API
• Cost values should reflect current pricing from the provider
• Context length should account for both input and output tokens
• Temperature ranges may vary by provider (most use 0-2, some use 0-1)

tip

Best practices for upserting seed models:

• Batch updates: Include all models in a single request to maintain consistency
• Accurate pricing: Regularly update cost information to reflect provider changes
• Version tracking: Use descriptive ModelGoodName values to differentiate versions
• Validation: Test model configurations after upserting
• Documentation: Keep records of when and why seed data was updated

Temperature guidelines by use case:

• 0.0-0.3: Factual, deterministic responses (documentation, data extraction)
• 0.4-0.7: Balanced creativity and consistency (general chat, Q&A)
• 0.8-1.0: Creative content (writing, brainstorming)
• 1.1-2.0: Highly experimental (not recommended for most use cases)

Cost calculation example:

• Input: 1M tokens at $10/M = $10.00
• Output: 500K tokens at $30/M = $15.00
• Total cost: $25.00 for the interaction

Response

Success Response (200 OK)

Returns an object confirming the upsert operation.

{
  "success": true,
  "provider": "openai",
  "models_updated": 2,
  "timestamp": "2025-01-12T10:30:00Z"
}

Response Fields

Field	Type	Description
-	object	Flexible response object (additionalProp1: {}) containing operation details. Typically includes success status, provider name, and count of models updated.

note

Response Object Structure

The actual response structure may vary but typically includes:

• Operation success status
• Provider name that was updated
• Number of models created or updated
• Timestamp of the operation
• Any warnings or validation messages

Error Response (422 Unprocessable Entity)

Returns validation error details when the request body is invalid or the provider doesn't exist.

{
  "detail": [
    {
      "loc": [
        "body",
        0,
        "MaxTokens"
      ],
      "msg": "MaxTokens must be a positive integer",
      "type": "value_error.number.not_gt"
    }
  ]
}

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 index).
msg	string	Human-readable error message.
type	string	Error type identifier.

Common Error Scenarios

Error Type	Description	Resolution
Provider Not Found	The specified provider doesn't exist in the system	Create the provider first or verify provider name
Invalid Model Format	Model identifier contains invalid characters	Use alphanumeric characters, hyphens, and underscores only
Missing Required Field	One or more required fields are missing	Ensure all required fields are included in each model object
Invalid Number Range	MaxTokens, DefaultTemp, or costs are out of valid range	Check value constraints (e.g., DefaultTemp typically 0-2)
Duplicate Model Names	Multiple models with the same Model identifier	Ensure Model identifiers are unique within the array
Invalid Cost Values	Negative or non-numeric cost values	Provide valid positive numbers for cost fields

Error Codes

Status Code	Description	Response Type
200	OK - Models upserted successfully	Success
400	Bad Request - Invalid request format	Bad Request
404	Not Found - Provider doesn't exist	Not Found
422	Validation Error - Invalid model data	Unprocessable Entity
500	Internal Server Error - Failed to upsert models	Internal Server Error

warning

Important Considerations

• This operation replaces the entire models array for the provider
• Always include all models you want to keep, not just new ones
• Upserted seed data affects all users and projects in the system
• Changes take effect immediately and impact model selection UI
• Incorrect pricing data can lead to inaccurate cost estimates
• Test thoroughly in a non-production environment first
• Keep audit logs of seed data changes
• Coordinate with team before modifying shared seed data
• Consider backwards compatibility when updating model identifiers
• Provider API changes may require seed data updates