Language Models

Call the platform's language models (LLM) for text conversation, content generation, function calling, and multimodal understanding. The gateway auto-detects the inbound request format and is compatible with OpenAI Chat Completions, Claude Messages, and native Gemini formats — just point your existing SDK's baseURL at https://apione.org, no adapter layer needed.

• Base URL: https://apione.org
• Primary endpoint (recommended): POST /v1/chat/completions (OpenAI format)
• Compatible endpoints: POST /v1/messages (Claude Messages format) / POST /v1beta/models/{model}:generateContent (native Gemini format)
• Authentication: Authorization: Bearer sk-xxxxxxxxxxxx (Claude format uses the x-api-key header; Gemini format uses the x-goog-api-key header or a ?key= query param)

Unlike image/video generation, chat completion is a synchronous API: non-streaming requests return the full result in one response, no task_id polling needed. Set stream: true for SSE streaming.

Three Request Formats

Request Example (OpenAI format)

curl --request POST \
  --url https://apione.org/v1/chat/completions \
  --header 'Authorization: Bearer sk-xxxxxxxxxxxx' \
  --header 'Content-Type: application/json' \
  --data '{
    "model": "gpt-5.4",
    "messages": [
      { "role": "system", "content": "You are a helpful assistant" },
      { "role": "user", "content": "Write an opening line for a sci-fi novel" }
    ],
    "temperature": 0.7,
    "max_tokens": 2000,
    "stream": false
  }'

Key Parameters: messages, stream, temperature, top_p, max_tokens / max_completion_tokens, reasoning_effort (low/medium/high, some reasoning models), tools / tool_choice (function calling), response_format (e.g. {"type":"json_object"} for JSON mode), n, seed, stop, frequency_penalty / presence_penalty

Vision: pass an image_url item inside messages[].content (same as OpenAI Vision format) — public URLs and Base64 Data URIs are both supported, subject to the selected model's multimodal capability.

Streaming (stream: true)

With "stream": true, the response is text/event-stream; each data: chunk is an incremental JSON delta identical to OpenAI's streaming format, ending with data: [DONE].

Supported Models

Billed by token usage. Currently verified models in active use:

The gateway is a multi-channel architecture and may also route to Claude, DeepSeek, and other model channels. The actual available set depends on live site configuration — query GET /v1/models for the authoritative, up-to-date list.

List Available Models

GET https://apione.org/v1/models

Returns OpenAI format by default. Send x-api-key + anthropic-version headers for Claude format, or x-goog-api-key / ?key= for Gemini format.

Common Error Codes