Skip to main content
kombify AI exposes a REST API for programmatic interaction. All endpoints require JWT authentication via Zitadel OIDC.

​
Base URL

ModeURL
SaaShttps://api.kombify.io/ai/v1
Self-hostedhttp://your-server:5270/api/v1

​
Authentication

All requests require a valid JWT token in the Authorization header: 
curl -H "Authorization: Bearer <your-jwt-token>" \
  https://api.kombify.io/ai/v1/chat

​
Endpoints

​
Chat

​
POST /api/v1/chat

Send a message and receive an AI response. 
{
  "message": "What is the status of my servers?",
  "companion": "smart-home",
  "model": "gpt-4o",
  "stream": true
}
Response: Server-Sent Events (SSE) stream when stream: true, or JSON response.

​
Companions

​
GET /api/v1/companions

List available Companions and their configuration.

​
PUT /api/v1/companions/:id/config

Update a Companion’s configuration.

​
Guard

​
GET /api/v1/guard/status

Get Guard’s current status, trust level, and recent activity.

​
POST /api/v1/guard/config

Update Guard configuration (trust level, intervals, targets).

​
Models

​
GET /api/v1/models

List available models and their status.

​
Usage

​
GET /api/v1/usage

Get token usage statistics and cost estimates. 
{
  "period": "2026-03",
  "total_tokens": 1250000,
  "estimated_cost_usd": 3.45,
  "by_provider": {
    "openai": { "tokens": 800000, "cost": 1.20 },
    "anthropic": { "tokens": 450000, "cost": 2.25 }
  }
}
This is a simplified reference. For the full OpenAPI specification, see the self-hosted AI portal at /api/docs.

​
Further reading

Configuration

All configuration options

Architecture

How the API fits into the overall system