OpenAI Compatibility API

GoingMerry exposes an OpenAI-compatible API layer directly on the daemon port. This allows tools designed for the OpenAI API to swap in GoingMerry as a local backend.

• Base Endpoint: http://localhost:11434/v1
• Authentication: Set the Authorization header to any dummy key (e.g. Bearer merry).

---

1. Chat Completions

Generate completions matching the OpenAI JSON schema.

• Endpoint: POST /v1/chat/completions
• Supported Parameters:
  • model (string, required): Local model tag.
  • messages (array of objects): Chat message thread history.
  • stream (boolean): Stream response delta chunks.
  • temperature (float): Creativity control.
  • max_tokens (integer): Output token limits (maps to num_predict).

curl Example

curl http://localhost:11434/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer merry" \
  -d '{
    "model": "gemma4",
    "messages": [
      {"role": "user", "content": "Hello!"}
    ],
    "stream": false
  }'

Response Payload

{
  "id": "chatcmpl-103ab28f9a0c",
  "object": "chat.completion",
  "created": 1780730100,
  "model": "gemma4",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Hello! How can I help you today?"
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 12,
    "completion_tokens": 8,
    "total_tokens": 20
  }
}

---

2. Generate Embeddings

Generate vector representations of text matching the OpenAI embeddings schema.

• Endpoint: POST /v1/embeddings

curl Example

curl http://localhost:11434/v1/embeddings \
  -H "Content-Type: application/json" \
  -d '{
    "model": "nomic-embed-text",
    "input": "Vectorize this text"
  }'

Response Payload

{
  "object": "list",
  "data": [
    {
      "object": "embedding",
      "index": 0,
      "embedding": [0.01723901, -0.0823901, 0.1023812, ...]
    }
  ],
  "model": "nomic-embed-text",
  "usage": {
    "prompt_tokens": 4,
    "total_tokens": 4
  }
}

---