API Reference

Contents

1. Basics

• Base URL: http://localhost:<port> (default port 11434). While on Wi-Fi, a LAN URL (http://<device-ip>:<port>) is also available on the same port.
• Shared port: The API and the WebUI are served on the same port. GET requests outside API paths fall through to the WebUI.
• Authentication: There is no API-key auth (the Authorization header is ignored). Intended for same-device or local-network use.
• CORS: All responses include Access-Control-Allow-Origin: *, and OPTIONS preflight returns 204.
• Content types: Request bodies are JSON (application/json). Non-streaming responses are JSON; streaming responses use NDJSON (application/x-ndjson) for /api/* and SSE (text/event-stream) for /v1/*.
• Model selection: model is a configuration (profile) name. List profiles via /api/tags or /v1/models.

2. Compatibility summary

For the range that typical chat / generation clients use, the API is compatible with both the Ollama and OpenAI specifications. Because the app is built around a single on-device model and a single generation slot, some management endpoints and statistics fields are not provided.

If you need token-usage style statistics, enable "Show Performance Metrics" in Output Settings. Token counts, processing time, and speed are appended after generation (in the final delta for OpenAI, as an extra chunk for Ollama).

3. Common behavior

3-1. Concurrency and busy queue

• There is a single inference slot (single generation). Additional requests during generation are queued (up to 10) and wait up to 60 seconds.
• Queue overflow or wait timeout returns 503.
• While a model reset is in progress, new requests are rejected and queued requests are aborted (503).

3-2. Error responses

/api/* endpoints return plain JSON / text errors, while /v1/* endpoints return an OpenAI-style error object.

// /v1/* error shape
{
  "error": {
    "message": "No messages provided",
    "type": "invalid_request_error",
    "code": 400
  }
}

4. POST/api/generate

Single-prompt generation (Ollama generate compatible). Streaming (NDJSON) by default.

Request

Non-streaming response

curl http://localhost:11434/api/generate -d '{
  "model": "my-profile",
  "prompt": "What is the capital of Japan?",
  "stream": false
}'

{
  "model": "my-profile",
  "created_at": "2026-06-14T03:10:00Z",
  "response": "The capital of Japan is Tokyo.",
  "done": true
}

Streaming response (NDJSON)

One JSON object per line. Each chunk has "done": false with a partial response; a final chunk with "done": true closes the stream.

{"model":"my-profile","created_at":"...","response":"The","done":false}
{"model":"my-profile","created_at":"...","response":" capital is","done":false}
{"model":"my-profile","created_at":"...","response":"","done":true}

5. POST/api/chat

Conversation generation from a messages array (Ollama chat compatible). Streaming (NDJSON) by default.

Request

Non-streaming response

curl http://localhost:11434/api/chat -d '{
  "model": "my-profile",
  "stream": false,
  "messages": [
    {"role": "system", "content": "You are a concise assistant."},
    {"role": "user", "content": "How tall is Mount Fuji?"}
  ]
}'

{
  "model": "my-profile",
  "created_at": "2026-06-14T03:10:00Z",
  "message": { "role": "assistant", "content": "3,776 m." },
  "done": true
}

Streaming response (NDJSON)

{"model":"my-profile","created_at":"...","message":{"role":"assistant","content":"3,776"},"done":false}
{"model":"my-profile","created_at":"...","message":{"role":"assistant","content":" m."},"done":false}
{"model":"my-profile","created_at":"...","message":{"role":"assistant","content":""},"done":true}

6. GETPOST/api/tags

List of available models (profiles). Ollama tags compatible.

curl http://localhost:11434/api/tags

{
  "models": [
    {
      "name": "my-profile",
      "model": "my-profile",
      "modified_at": "2026-06-14T03:10:00Z",
      "size": 0,
      "details": {
        "format": "gguf",
        "family": "llama",
        "parameter_size": "unknown",
        "quantization_level": "unknown"
      }
    }
  ]
}

Fields such as size and parameter_size cannot be determined on-device, so they use fixed values (0 / "unknown"). Use /v1/models for richer status and modality information.

7. POST/v1/chat/completions

OpenAI Chat Completions compatible. Streaming (SSE) by default. Point the official OpenAI SDK's base_url at this server to use it.

Request

Non-streaming response

curl http://localhost:11434/v1/chat/completions -d '{
  "model": "my-profile",
  "stream": false,
  "messages": [
    {"role": "user", "content": "Say hello in Japanese."}
  ]
}'

{
  "object": "chat.completion",
  "model": "my-profile",
  "choices": [
    {
      "index": 0,
      "message": { "role": "assistant", "content": "Konnichiwa!" },
      "finish_reason": "stop"
    }
  ]
}

Note: responses do not include id / created / usage. Clients that require these will see them as missing.

Streaming response (SSE)

First a chunk with delta.role="assistant", then chunks carrying delta.content, and finally a chunk with finish_reason:"stop" followed by data: [DONE].

data: {"object":"chat.completion.chunk","model":"my-profile","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}

data: {"object":"chat.completion.chunk","model":"my-profile","choices":[{"index":0,"delta":{"content":"Konni"},"finish_reason":null}]}

data: {"object":"chat.completion.chunk","model":"my-profile","choices":[{"index":0,"delta":{"content":"chiwa!"},"finish_reason":null}]}

data: {"object":"chat.completion.chunk","model":"my-profile","choices":[{"index":0,"delta":{},"finish_reason":"stop"}]}

data: [DONE]

8. GET/v1/models & /models

Returns the OpenAI-compatible model list plus a llama.cpp WebUI-oriented models array (status and modalities).

curl http://localhost:11434/v1/models

{
  "object": "list",
  "data": [
    {
      "id": "my-profile",
      "name": "my-profile",
      "object": "model",
      "owned_by": "llamacpp",
      "created": 1781000000,
      "in_cache": true,
      "path": "/data/.../my-model.gguf",
      "status": { "value": "loaded" },
      "tags": []
    }
  ],
  "models": [
    {
      "name": "my-profile",
      "model": "my-profile",
      "modified_at": "2026-06-14T03:10:00Z",
      "size": 4100000000,
      "capabilities": ["multimodal"],
      "modalities": { "vision": true, "audio": false },
      "details": { "format": "gguf", "family": "llama", "parameter_size": "unknown", "quantization_level": "unknown" }
    }
  ]
}

data[] is OpenAI-compatible; models[] is llama.cpp WebUI-compatible. modalities reflects actual support for loaded models, or values inferred from the file when not loaded.

9. GET/props, /slots, /health

Endpoints mainly used to initialize the bundled (llama.cpp-style) WebUI. Use ?model=<profile-name> to target a profile.

• /props: Returns default_generation_settings (n_ctx and sampling values), chat_template, modalities, model_path, build_info, and webui_settings (system message, Think display, shared MCP / Function Definitions). total_slots is 1.
• /slots: A slot array (one element) with n_ctx, params (sampling values), and is_processing.
• /health, /v1/health: {"status":"ok","role":...,"webui":true}.

curl "http://localhost:11434/props?model=my-profile"

10. Sampling parameters

For /v1/chat/completions, including the following keys in the request body overrides the configured values for that request (unspecified fields fall back to the profile values).

Setting n_predict: 0 performs pre-encode only and returns an empty response early (useful for prompt-evaluation warm-up).

11. Multimodal inputs

/api/chat and /v1/chat/completions accept images and audio by passing content as a parts array. The loaded model must support vision / audio; otherwise the request returns 400.

curl http://localhost:11434/v1/chat/completions -d '{
  "model": "my-vision-profile",
  "stream": false,
  "messages": [
    {
      "role": "user",
      "content": [
        {"type": "text", "text": "Describe this image."},
        {"type": "image_url", "image_url": {"url": "data:image/jpeg;base64,/9j/4AAQ..."}}
      ]
    }
  ]
}'

// Audio input (wav / mp3)
{
  "role": "user",
  "content": [
    {"type": "text", "text": "Transcribe this audio."},
    {"type": "input_audio", "input_audio": {"data": "UklGR...", "format": "wav"}}
  ]
}

12. Tools / Function Calling

Passing tools in the request, or enabling shared MCP / Function Definitions in the app's MCP Settings, switches to the internal tool-execution loop. This works on /api/generate, /api/chat, and /v1/chat/completions.

• tools is an OpenAI-compatible array. Passing a non-array returns 400.
• tool_choice and parallel_tool_calls are accepted.
• When shared MCP / Function Definitions are enabled, the same tool settings apply to the main prompt input and to each API.
• Responses may include tool_calls returned directly by the model, plus calls extracted from in-text tool-call markers or reasoning_content.

// OpenAI-compatible tool example (part of a non-streaming response)
{
  "object": "chat.completion",
  "model": "my-profile",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "",
        "tool_calls": [
          {
            "id": "call_1",
            "type": "function",
            "function": { "name": "get_weather", "arguments": "{\"city\":\"Tokyo\"}" }
          }
        ]
      },
      "finish_reason": "tool_calls"
    }
  ]
}

13. Client examples

OpenAI Python SDK

from openai import OpenAI

client = OpenAI(
    base_url="http://localhost:11434/v1",
    api_key="not-needed",  # auth is not required (any string works)
)

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

OpenAI-compatible streaming (Python)

stream = client.chat.completions.create(
    model="my-profile",
    messages=[{"role": "user", "content": "Write one haiku"}],
    stream=True,
)
for chunk in stream:
    delta = chunk.choices[0].delta
    if delta.content:
        print(delta.content, end="", flush=True)

Ollama-compatible (curl / NDJSON)

curl -N http://localhost:11434/api/chat -d '{
  "model": "my-profile",
  "messages": [{"role": "user", "content": "Introduce yourself"}]
}'

JavaScript (fetch / SSE)

const res = await fetch("http://localhost:11434/v1/chat/completions", {
  method: "POST",
  headers: { "Content-Type": "application/json" },
  body: JSON.stringify({
    model: "my-profile",
    messages: [{ role: "user", content: "Hello" }],
    stream: true,
  }),
});

const reader = res.body.getReader();
const decoder = new TextDecoder();
while (true) {
  const { value, done } = await reader.read();
  if (done) break;
  for (const line of decoder.decode(value).split("\n")) {
    if (!line.startsWith("data: ")) continue;
    const data = line.slice(6);
    if (data === "[DONE]") break;
    const json = JSON.parse(data);
    process.stdout.write(json.choices[0].delta.content ?? "");
  }
}

This page is based on the current Android implementation (OllamaApiServer). Response examples are formatted for clarity; actual field sets and defaults may vary by app version.