LM Studio 本地 LLM
API 与 SDK 全景图

# 启动服务器（默认端口 1234）
lms server start

# 也可以指定端口
lms server start --port 1234

// npm install @lmstudio/sdk
import { LMStudioClient } from "@lmstudio/sdk";

const client = new LMStudioClient();
const model  = await client.llm.model("openai/gpt-oss-20b");
const result = await model.respond("Who are you, and what can you do?");

console.info(result.content);

# pip install lmstudio
import lmstudio as lms

with lms.Client() as client:
    model  = client.llm.model("openai/gpt-oss-20b")
    result = model.respond("Who are you, and what can you do?")
    print(result)

curl http://localhost:1234/api/v1/chat \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $LM_API_TOKEN" \
  -d '{
    "model": "openai/gpt-oss-20b",
    "input": "Who are you, and what can you do?"
  }'

curl -X POST http://localhost:1234/api/v1/chat \
  -H "Authorization: Bearer $LM_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{ "model": "ibm/granite-4-micro", "input": "Open lmstudio.ai" }'

curl http://localhost:1234/api/v1/chat \
  -H "Authorization: Bearer $LM_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "ibm/granite-4-micro",
    "input": "My favorite color is blue."
  }'

{
  "model_instance_id": "ibm/granite-4-micro",
  "output": [
    { "type": "message", "content": "That's great! Blue is a beautiful color…" }
  ],
  "response_id": "resp_abc123xyz…"
}

curl http://localhost:1234/api/v1/chat \
  -H "Authorization: Bearer $LM_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "ibm/granite-4-micro",
    "input": "What color did I just mention?",
    "previous_response_id": "resp_abc123xyz…"
  }'

curl http://localhost:1234/api/v1/chat \
  -H "Authorization: Bearer $LM_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{ "model": "ibm/granite-4-micro", "input": "Tell me a joke.", "store": false }'

"stats": {
  "input_tokens": 646,
  "total_output_tokens": 586,
  "reasoning_output_tokens": 0,
  "tokens_per_second": 29.75,            // 生成速度
  "time_to_first_token_seconds": 1.088,    // 首 token 延迟 (TTFT)
  "model_load_time_seconds": 2.656          // 仅当本次需加载模型时出现
}

"input": [
  { "type": "message", "content": "What's in this image?" },
  { "type": "image",   "data_url": "data:image/jpeg;base64,…" }
]

from openai import OpenAI

client = OpenAI(
    base_url="http://localhost:1234/v1",   # 唯一改动
    api_key="lm-studio"                      # 任意占位符即可
)
# …其余代码与调用 OpenAI 完全相同…

- curl https://api.openai.com/v1/chat/completions \
+ curl http://localhost:1234/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
      "model": "在此填入 LM Studio 的模型标识",
      "messages": [{"role": "user", "content": "Say this is a test!"}],
      "temperature": 0.7
   }'

from openai import OpenAI
import json

client = OpenAI(base_url="http://localhost:1234/v1", api_key="lm-studio")

character_schema = {
    "type": "json_schema",
    "json_schema": {
        "name": "characters",
        "schema": {
            "type": "object",
            "properties": {
                "characters": {
                    "type": "array",
                    "items": {
                        "type": "object",
                        "properties": {
                            "name":        {"type": "string"},
                            "occupation":  {"type": "string"},
                            "personality": {"type": "string"},
                            "background":  {"type": "string"}
                        },
                        "required": ["name","occupation","personality","background"]
                    },
                    "minItems": 1
                }
            },
            "required": ["characters"]
        }
    }
}

response = client.chat.completions.create(
    model="your-model",
    messages=[{"role":"user","content":"Create 1-3 fictional characters"}],
    response_format=character_schema,
)

# content 是 JSON 字符串，需要自行解析
results = json.loads(response.choices[0].message.content)
print(json.dumps(results, indent=2))

[
  {
    "type": "function",
    "function": {
      "name": "get_delivery_date",
      "description": "Get the delivery date for a customer's order",
      "parameters": {
        "type": "object",
        "properties": { "order_id": { "type": "string" } },
        "required": ["order_id"]
      }
    }
  }
]

# 1) 第一次调用，带上 tools
response = client.chat.completions.create(model=model, messages=messages, tools=tools)

# 2) 解析模型请求的工具调用，执行真实函数
tool_call = response.choices[0].message.tool_calls[0]
arguments = json.loads(tool_call.function.arguments)
delivery_date = get_delivery_date(arguments["order_id"])

# 3) 把"助手的工具调用"和"工具结果"都追加进 messages
messages.append({"role":"assistant", "tool_calls":[…]})
messages.append({
    "role": "tool",
    "content": json.dumps({"delivery_date": delivery_date}),
    "tool_call_id": tool_call.id
})

# 4) 再次调用（这次不带 tools），拿到自然语言最终回答
final = client.chat.completions.create(model=model, messages=messages)
print(final.choices[0].message.content)
# → "Your order #1017 is scheduled for delivery on November 19, 2024…"

curl http://localhost:1234/api/v1/chat \
  -H "Authorization: Bearer $LM_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "ibm/granite-4-micro",
    "input": "What is the top trending model on hugging face?",
    "integrations": [
      {
        "type": "ephemeral_mcp",
        "server_label": "huggingface",
        "server_url": "https://huggingface.co/mcp",
        "allowed_tools": ["model_search"]
      }
    ],
    "context_length": 8000
  }'

curl http://localhost:1234/api/v1/chat \
  -H "Authorization: Bearer $LM_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "ibm/granite-4-micro",
    "input": "Open lmstudio.ai",
    "integrations": [
      { "type": "plugin", "id": "mcp/playwright", "allowed_tools": ["browser_navigate"] }
    ],
    "context_length": 8000
  }'

curl -fsSL https://lmstudio.ai/install.sh | bash

irm https://lmstudio.ai/install.ps1 | iex

lms daemon up          # 启动守护进程
lms get <model>        # 下载模型
lms server start       # 启动本地服务器
lms chat               # 打开交互式会话
lms load               # 加载一个模型进内存
lms log stream         # 查看实时日志（排错神器）