API Reference

Chat Completions (LLM)

OpenAI-compatible chat completions API — call qyma-chat or qyma-logic with your External API key.

Overview

SkyLight Chat exposes an OpenAI-compatible chat completions endpoint so you can point the OpenAI SDK (or any compatible client) at SkyLight and use the branded models qyma-chat and qyma-logic.

Same authentication as the rest of External API v1 (X-Api-Key or Authorization: Bearer)
Responses follow the OpenAI chat.completion shape
Usage is metered to your account (same AI usage pipeline as the dashboard)
Streaming is not supported in v1 (stream: true returns 400)
Available models: qyma-chat (default), qyma-logic

Base URL

https://dashboard.skylightchat.com/api/v1

OpenAI SDK example:

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_EXTERNAL_API_KEY",
    base_url="https://dashboard.skylightchat.com/api/v1",
)

resp = client.chat.completions.create(
    model="qyma-chat",
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

List models

GET /api/v1/models

{
  "object": "list",
  "data": [
    {
      "id": "qyma-chat",
      "object": "model",
      "created": 1754006400,
      "owned_by": "qyma"
    },
    {
      "id": "qyma-logic",
      "object": "model",
      "created": 1754265600,
      "owned_by": "qyma"
    }
  ]
}

Get a model

GET /api/v1/models/qyma-chat

GET /api/v1/models/qyma-logic

Unknown models return 404 with an OpenAI-style error object.

Create a chat completion

POST /api/v1/chat/completions
Content-Type: application/json

Field	Type	Required	Notes
`model`	string	No	`qyma-chat` (default) or `qyma-logic`
`messages`	array	Yes	OpenAI chat messages (`role` + `content`)
`temperature`	number	No	Sampling temperature
`max_tokens` / `max_completion_tokens`	integer	No	Max output tokens
`response_format`	object	No	e.g. `{ "type": "json_object" }` (supported on `qyma-logic`)
`tools`	array	No	Tool definitions (OpenAI format)
`tool_choice`	string\|object	No	Tool choice
`stream`	boolean	No	Must be `false` or omitted — streaming not supported

curl https://dashboard.skylightchat.com/api/v1/chat/completions \
  -H "X-Api-Key: ••••••••••••" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qyma-logic",
    "messages": [
      {"role": "user", "content": "Summarize our return policy in one sentence."}
    ]
  }'

Response (200)

{
  "id": "chatcmpl-…",
  "object": "chat.completion",
  "model": "qyma-logic",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "…"
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 12,
    "completion_tokens": 28,
    "total_tokens": 40
  }
}

Errors

Errors use the OpenAI error envelope (not the usual { success, message } wrapper):

{
  "error": {
    "message": "Invalid model. Use qyma-chat or qyma-logic.",
    "type": "invalid_request_error",
    "code": "model_not_found"
  }
}

Code	HTTP	When
`model_not_found`	400 / 404	Unknown model id
`stream_not_supported`	400	`stream: true`
`invalid_messages`	400	Missing / empty `messages`
(server)	502	Generation failed — retry later

Authentication & rate limits

Identical to the rest of External API v1:

Authorization: Bearer <api_key> or X-Api-Key: <api_key>
Plan must have API access enabled
120 requests / minute per API key

Generate a key in the dashboard under Settings → API.

Connect LinkedIn, search prospects, run invite and sequence campaigns, and manage Recruiter jobs via the API.

Messages

List conversations and send messages via WhatsApp, Telegram, Instagram, and LinkedIn.