GET /api/models

​

Authentication

​

Response

curl -X GET 'http://localhost:5079/api/models' \
  -H 'Authorization: Bearer YOUR_JWT_TOKEN'

const response = await fetch('http://localhost:5079/api/models', {
  headers: {
    'Authorization': `Bearer ${jwt}`
  }
});

const models = await response.json();

import requests

response = requests.get(
    'http://localhost:5079/api/models',
    headers={'Authorization': f'Bearer {jwt}'}
)

models = response.json()

{
  "items": [
    {
      "modelId": "model-uuid-1",
      "modelName": "llama-3.3-70b-versatile",
      "displayName": "Llama 3.3 70B Versatile",
      "providerName": "groq",
      "apiUrl": "https://api.groq.com/openai/v1/chat/completions",
      "status": "active"
    },
    {
      "modelId": "model-uuid-2",
      "modelName": "mixtral-8x7b-32768",
      "displayName": "Mixtral 8x7B",
      "providerName": "groq",
      "apiUrl": "https://api.groq.com/openai/v1/chat/completions",
      "status": "active"
    }
  ]
}

{
  "error": "Unauthorized"
}

{
  "success": false,
  "error": "Database query failed",
  "code": "MODELS_ERROR"
}

​

Side Effects

var rows = await _supabase.SelectAsync<JObject>(
    "ai_models",
    "model_id,model_name,provider_name,api_url,description,status,created_at",
    "status=eq.active&order=created_at.desc"
);

• ai_models table

• WHERE status = 'active'
• ORDER BY created_at DESC

​

Authorization

• Any authenticated user

• Global model list (not user-specific)
• All users see same models

​

Permissions

​

Response Structure

• Ordered by created_at DESC (newest models first)
• Hardcoded in query, not configurable

• Only status = 'active' models returned
• Inactive models excluded
• No additional filtering parameters

• Returns ALL active models
• Performance depends on number of active models

​

Edge Cases

1. No active models: Returns {"items": []} (empty array, Line 35)
2. Model missing description: displayName falls back to modelName (Line 39)
3. Null fields: Converted to null in JSON response (Lines 37-42)
4. Very large result set: No pagination, could be slow

​

Error Conditions

catch (Exception ex) {
    return StatusCode(500, new { error = ex.Message, code = "MODELS_ERROR" });
}

• All exceptions return 500
• Exception message exposed to client

​

Field Mapping

ai_models.model_id → modelId
ai_models.model_name → modelName
ai_models.description → displayName (fallback: model_name)
ai_models.provider_name → providerName
ai_models.api_url → apiUrl
ai_models.status → status

​

Behavioral Guarantees

• Caching behavior not enforced by server contract
• Real-time database query (no caching documented in controller)

• Models added/removed during request processing may not be reflected

• Partial results not enforced by server contract

​

Performance Characteristics

• ai_models.status should be indexed
• ai_models.created_at should be indexed for sorting

​

Use Cases

• Populate model dropdown in UI
• GetRandomModelAsync service queries this data
• GetTwoRandomModelsAsync for dual-chat

• Determine available providers
• Get API endpoints for model routing

• Show model names and descriptions to users

​

Status Values

• "inactive": Models disabled/deprecated
• "testing": Models in beta
• Behavior not enforced by server contract (database schema-dependent)

​

Provider Information

• Full provider API endpoint
• Used by chat providers for routing
• Exposed to authenticated clients