Mjara Docs
Architecture

API Layer

The API Layer provides RESTful endpoints built on the Mjara API framework. All endpoints follow the pattern:

/api/method/arif_ai.api.{endpoint}

Modules

ModulePurpose
chat.pyCore chat endpoints with context auto-summarization
conversations.pyConversation CRUD and history management
tools.pyTool discovery, execution, and invocation
onboarding.pyOnboarding workflow management
confirmations.pyHuman-in-the-loop confirmation handling
streaming.pyReal-time SSE streaming
settings.pyConfiguration management
files.pyFile attachment handling

Authentication

MethodDescription
Session AuthAutomatic for logged-in users
API Keyuser@example.com:api_secret
Bearer TokenOptional JWT support

Rate Limiting

  • Chat & Conversations: 100 requests/minute
  • Tool Executions: 50 requests/minute

Rate limit headers: X-RateLimit-Limit, X-RateLimit-Remaining, X-RateLimit-Reset

Response Format

All endpoints return a standard structure:

{
  "message": {
    "success": true,
    "data": {},
    "tokens_used": {
      "input": 150,
      "output": 50,
      "total": 200
    }
  }
}

For full endpoint documentation, see the API Reference.

On this page