---
title: "Instructor"
sidebar_label: "Instructor"
description: "使用 Pydantic 验证从 LLM 响应中提取结构化数据，自动重试失败的提取，以类型安全方式解析复杂 JSON，并使用 Instructor 流式传输部分结果——经过实战检验的结构化输出库"
---

{/* This page is auto-generated from the skill's SKILL.md by website/scripts/generate-skill-docs.py. Edit the source SKILL.md, not this page. */}

# Instructor

使用 Pydantic 验证从 LLM 响应中提取结构化数据，自动重试失败的提取，以类型安全方式解析复杂 JSON，并使用 Instructor 流式传输部分结果——经过实战检验的结构化输出库

## Skill 元数据

| | |
|---|---|
| 来源 | 可选 — 通过 `hermes skills install official/mlops/instructor` 安装 |
| 路径 | `optional-skills/mlops/instructor` |
| 版本 | `1.0.0` |
| 作者 | Orchestra Research |
| 许可证 | MIT |
| 依赖项 | `instructor`, `pydantic`, `openai`, `anthropic` |
| 平台 | linux, macos, windows |
| 标签 | `Prompt Engineering`, `Instructor`, `Structured Output`, `Pydantic`, `Data Extraction`, `JSON Parsing`, `Type Safety`, `Validation`, `Streaming`, `OpenAI`, `Anthropic` |

## 参考：完整 SKILL.md

:::info
以下是 Hermes 在触发此 skill 时加载的完整 skill 定义。这是 skill 激活时 agent 所看到的指令内容。
:::

# Instructor：结构化 LLM 输出

## 何时使用此 Skill

在以下情况下使用 Instructor：
- **从 LLM 响应中可靠地提取结构化数据**
- **根据 Pydantic schema 自动验证输出**
- **通过自动错误处理重试失败的提取**
- **以类型安全和验证方式解析复杂 JSON**
- **流式传输部分结果**以进行实时处理
- **以一致的 API 支持多个 LLM 提供商**

**GitHub Stars**：15,000+｜**实战检验**：100,000+ 开发者

## 安装

```bash
# 基础安装
pip install instructor

# 指定提供商
pip install "instructor[anthropic]"  # Anthropic Claude
pip install "instructor[openai]"     # OpenAI
pip install "instructor[all]"        # 所有提供商
```

## 快速开始

### 基础示例：提取用户数据

```python
import instructor
from pydantic import BaseModel
from anthropic import Anthropic

# Define output structure
class User(BaseModel):
    name: str
    age: int
    email: str

# Create instructor client
client = instructor.from_anthropic(Anthropic())

# Extract structured data
user = client.messages.create(
    model="claude-sonnet-4-5-20250929",
    max_tokens=1024,
    messages=[{
        "role": "user",
        "content": "John Doe is 30 years old. His email is john@example.com"
    }],
    response_model=User
)

print(user.name)   # "John Doe"
print(user.age)    # 30
print(user.email)  # "john@example.com"
```

### 使用 OpenAI

```python
from openai import OpenAI

client = instructor.from_openai(OpenAI())

user = client.chat.completions.create(
    model="gpt-4o-mini",
    response_model=User,
    messages=[{"role": "user", "content": "Extract: Alice, 25, alice@email.com"}]
)
```

## 核心概念

### 1. 响应模型（Pydantic）

响应模型定义 LLM 输出的结构和验证规则。

#### 基础模型

```python
from pydantic import BaseModel, Field

class Article(BaseModel):
    title: str = Field(description="Article title")
    author: str = Field(description="Author name")
    word_count: int = Field(description="Number of words", gt=0)
    tags: list[str] = Field(description="List of relevant tags")

article = client.messages.create(
    model="claude-sonnet-4-5-20250929",
    max_tokens=1024,
    messages=[{
        "role": "user",
        "content": "Analyze this article: [article text]"
    }],
    response_model=Article
)
```

**优势：**
- 使用 Python 类型提示保证类型安全
- 自动验证（word_count > 0）
- 通过 Field 描述实现自文档化
- IDE 自动补全支持

#### 嵌套模型

```python
class Address(BaseModel):
    street: str
    city: str
    country: str

class Person(BaseModel):
    name: str
    age: int
    address: Address  # Nested model

person = client.messages.create(
    model="claude-sonnet-4-5-20250929",
    max_tokens=1024,
    messages=[{
        "role": "user",
        "content": "John lives at 123 Main St, Boston, USA"
    }],
    response_model=Person
)

print(person.address.city)  # "Boston"
```

#### 可选字段

```python
from typing import Optional

class Product(BaseModel):
    name: str
    price: float
    discount: Optional[float] = None  # Optional
    description: str = Field(default="No description")  # Default value

# LLM doesn't need to provide discount or description
```

#### 使用枚举约束值

```python
from enum import Enum

class Sentiment(str, Enum):
    POSITIVE = "positive"
    NEGATIVE = "negative"
    NEUTRAL = "neutral"

class Review(BaseModel):
    text: str
    sentiment: Sentiment  # Only these 3 values allowed

review = client.messages.create(
    model="claude-sonnet-4-5-20250929",
    max_tokens=1024,
    messages=[{
        "role": "user",
        "content": "This product is amazing!"
    }],
    response_model=Review
)

print(review.sentiment)  # Sentiment.POSITIVE
```

### 2. 验证

Pydantic 自动验证 LLM 输出。若验证失败，Instructor 会自动重试。

#### 内置验证器

```python
from pydantic import Field, EmailStr, HttpUrl

class Contact(BaseModel):
    name: str = Field(min_length=2, max_length=100)
    age: int = Field(ge=0, le=120)  # 0 <= age <= 120
    email: EmailStr  # Validates email format
    website: HttpUrl  # Validates URL format

# If LLM provides invalid data, Instructor retries automatically
```

#### 自定义验证器

```python
from pydantic import field_validator

class Event(BaseModel):
    name: str
    date: str
    attendees: int

    @field_validator('date')
    def validate_date(cls, v):
        """Ensure date is in YYYY-MM-DD format."""
        import re
        if not re.match(r'\d{4}-\d{2}-\d{2}', v):
            raise ValueError('Date must be YYYY-MM-DD format')
        return v

    @field_validator('attendees')
    def validate_attendees(cls, v):
        """Ensure positive attendees."""
        if v < 1:
            raise ValueError('Must have at least 1 attendee')
        return v
```

#### 模型级验证

```python
from pydantic import model_validator

class DateRange(BaseModel):
    start_date: str
    end_date: str

    @model_validator(mode='after')
    def check_dates(self):
        """Ensure end_date is after start_date."""
        from datetime import datetime
        start = datetime.strptime(self.start_date, '%Y-%m-%d')
        end = datetime.strptime(self.end_date, '%Y-%m-%d')

        if end < start:
            raise ValueError('end_date must be after start_date')
        return self
```

### 3. 自动重试

当验证失败时，Instructor 会自动重试，并将错误反馈提供给 LLM。

```python
# Retries up to 3 times if validation fails
user = client.messages.create(
    model="claude-sonnet-4-5-20250929",
    max_tokens=1024,
    messages=[{
        "role": "user",
        "content": "Extract user from: John, age unknown"
    }],
    response_model=User,
    max_retries=3  # Default is 3
)

# If age can't be extracted, Instructor tells the LLM:
# "Validation error: age - field required"
# LLM tries again with better extraction
```

**工作原理：**
1. LLM 生成输出
2. Pydantic 进行验证
3. 若无效：将错误信息发回给 LLM
4. LLM 根据错误反馈重新尝试
5. 重复直至达到 max_retries 次数

### 4. 流式传输

流式传输部分结果以进行实时处理。

#### 流式传输部分对象

```python
from instructor import Partial

class Story(BaseModel):
    title: str
    content: str
    tags: list[str]

# Stream partial updates as LLM generates
for partial_story in client.messages.create_partial(
    model="claude-sonnet-4-5-20250929",
    max_tokens=1024,
    messages=[{
        "role": "user",
        "content": "Write a short sci-fi story"
    }],
    response_model=Story
):
    print(f"Title: {partial_story.title}")
    print(f"Content so far: {partial_story.content[:100]}...")
    # Update UI in real-time
```

#### 流式传输可迭代对象

```python
class Task(BaseModel):
    title: str
    priority: str

# Stream list items as they're generated
tasks = client.messages.create_iterable(
    model="claude-sonnet-4-5-20250929",
    max_tokens=1024,
    messages=[{
        "role": "user",
        "content": "Generate 10 project tasks"
    }],
    response_model=Task
)

for task in tasks:
    print(f"- {task.title} ({task.priority})")
    # Process each task as it arrives
```

## 提供商配置

### Anthropic Claude

```python
import instructor
from anthropic import Anthropic

client = instructor.from_anthropic(
    Anthropic(api_key="your-api-key")
)

# Use with Claude models
response = client.messages.create(
    model="claude-sonnet-4-5-20250929",
    max_tokens=1024,
    messages=[...],
    response_model=YourModel
)
```

### OpenAI

```python
from openai import OpenAI

client = instructor.from_openai(
    OpenAI(api_key="your-api-key")
)

response = client.chat.completions.create(
    model="gpt-4o-mini",
    response_model=YourModel,
    messages=[...]
)
```

### 本地模型（Ollama）

```python
from openai import OpenAI

# Point to local Ollama server
client = instructor.from_openai(
    OpenAI(
        base_url="http://localhost:11434/v1",
        api_key="ollama"  # Required but ignored
    ),
    mode=instructor.Mode.JSON
)

response = client.chat.completions.create(
    model="llama3.1",
    response_model=YourModel,
    messages=[...]
)
```

## 常用模式

### 模式 1：从文本中提取数据

```python
class CompanyInfo(BaseModel):
    name: str
    founded_year: int
    industry: str
    employees: int
    headquarters: str

text = """
Tesla, Inc. was founded in 2003. It operates in the automotive and energy
industry with approximately 140,000 employees. The company is headquartered
in Austin, Texas.
"""

company = client.messages.create(
    model="claude-sonnet-4-5-20250929",
    max_tokens=1024,
    messages=[{
        "role": "user",
        "content": f"Extract company information from: {text}"
    }],
    response_model=CompanyInfo
)
```

### 模式 2：分类

```python
class Category(str, Enum):
    TECHNOLOGY = "technology"
    FINANCE = "finance"
    HEALTHCARE = "healthcare"
    EDUCATION = "education"
    OTHER = "other"

class ArticleClassification(BaseModel):
    category: Category
    confidence: float = Field(ge=0.0, le=1.0)
    keywords: list[str]

classification = client.messages.create(
    model="claude-sonnet-4-5-20250929",
    max_tokens=1024,
    messages=[{
        "role": "user",
        "content": "Classify this article: [article text]"
    }],
    response_model=ArticleClassification
)
```

### 模式 3：多实体提取

```python
class Person(BaseModel):
    name: str
    role: str

class Organization(BaseModel):
    name: str
    industry: str

class Entities(BaseModel):
    people: list[Person]
    organizations: list[Organization]
    locations: list[str]

text = "Tim Cook, CEO of Apple, announced at the event in Cupertino..."

entities = client.messages.create(
    model="claude-sonnet-4-5-20250929",
    max_tokens=1024,
    messages=[{
        "role": "user",
        "content": f"Extract all entities from: {text}"
    }],
    response_model=Entities
)

for person in entities.people:
    print(f"{person.name} - {person.role}")
```

### 模式 4：结构化分析

```python
class SentimentAnalysis(BaseModel):
    overall_sentiment: Sentiment
    positive_aspects: list[str]
    negative_aspects: list[str]
    suggestions: list[str]
    score: float = Field(ge=-1.0, le=1.0)

review = "The product works well but setup was confusing..."

analysis = client.messages.create(
    model="claude-sonnet-4-5-20250929",
    max_tokens=1024,
    messages=[{
        "role": "user",
        "content": f"Analyze this review: {review}"
    }],
    response_model=SentimentAnalysis
)
```

### 模式 5：批量处理

```python
def extract_person(text: str) -> Person:
    return client.messages.create(
        model="claude-sonnet-4-5-20250929",
        max_tokens=1024,
        messages=[{
            "role": "user",
            "content": f"Extract person from: {text}"
        }],
        response_model=Person
    )

texts = [
    "John Doe is a 30-year-old engineer",
    "Jane Smith, 25, works in marketing",
    "Bob Johnson, age 40, software developer"
]

people = [extract_person(text) for text in texts]
```

## 高级特性

### 联合类型

```python
from typing import Union

class TextContent(BaseModel):
    type: str = "text"
    content: str

class ImageContent(BaseModel):
    type: str = "image"
    url: HttpUrl
    caption: str

class Post(BaseModel):
    title: str
    content: Union[TextContent, ImageContent]  # Either type

# LLM chooses appropriate type based on content
```

### 动态模型

```python
from pydantic import create_model

# Create model at runtime
DynamicUser = create_model(
    'User',
    name=(str, ...),
    age=(int, Field(ge=0)),
    email=(EmailStr, ...)
)

user = client.messages.create(
    model="claude-sonnet-4-5-20250929",
    max_tokens=1024,
    messages=[...],
    response_model=DynamicUser
)
```

### 自定义模式

```python
# For providers without native structured outputs
client = instructor.from_anthropic(
    Anthropic(),
    mode=instructor.Mode.JSON  # JSON mode
)

# Available modes:
# - Mode.ANTHROPIC_TOOLS (recommended for Claude)
# - Mode.JSON (fallback)
# - Mode.TOOLS (OpenAI tools)
```

### 上下文管理

```python
# Single-use client
with instructor.from_anthropic(Anthropic()) as client:
    result = client.messages.create(
        model="claude-sonnet-4-5-20250929",
        max_tokens=1024,
        messages=[...],
        response_model=YourModel
    )
    # Client closed automatically
```

## 错误处理

### 处理验证错误

```python
from pydantic import ValidationError

try:
    user = client.messages.create(
        model="claude-sonnet-4-5-20250929",
        max_tokens=1024,
        messages=[...],
        response_model=User,
        max_retries=3
    )
except ValidationError as e:
    print(f"Failed after retries: {e}")
    # Handle gracefully

except Exception as e:
    print(f"API error: {e}")
```

### 自定义错误信息

```python
class ValidatedUser(BaseModel):
    name: str = Field(description="Full name, 2-100 characters")
    age: int = Field(description="Age between 0 and 120", ge=0, le=120)
    email: EmailStr = Field(description="Valid email address")

    class Config:
        # Custom error messages
        json_schema_extra = {
            "examples": [
                {
                    "name": "John Doe",
                    "age": 30,
                    "email": "john@example.com"
                }
            ]
        }
```

## 最佳实践

### 1. 清晰的字段描述

```python
# ❌ Bad: Vague
class Product(BaseModel):
    name: str
    price: float

# ✅ Good: Descriptive
class Product(BaseModel):
    name: str = Field(description="Product name from the text")
    price: float = Field(description="Price in USD, without currency symbol")
```

### 2. 使用适当的验证

```python
# ✅ Good: Constrain values
class Rating(BaseModel):
    score: int = Field(ge=1, le=5, description="Rating from 1 to 5 stars")
    review: str = Field(min_length=10, description="Review text, at least 10 chars")
```

### 3. 在 prompt（提示词）中提供示例

```python
messages = [{
    "role": "user",
    "content": """Extract person info from: "John, 30, engineer"

Example format:
{
  "name": "John Doe",
  "age": 30,
  "occupation": "engineer"
}"""
}]
```

### 4. 对固定类别使用枚举

```python
# ✅ Good: Enum ensures valid values
class Status(str, Enum):
    PENDING = "pending"
    APPROVED = "approved"
    REJECTED = "rejected"

class Application(BaseModel):
    status: Status  # LLM must choose from enum
```

### 5. 优雅处理缺失数据

```python
class PartialData(BaseModel):
    required_field: str
    optional_field: Optional[str] = None
    default_field: str = "default_value"

# LLM only needs to provide required_field
```

## 与其他方案的对比

| 特性 | Instructor | 手动 JSON | LangChain | DSPy |
|---------|------------|-------------|-----------|------|
| 类型安全 | ✅ 是 | ❌ 否 | ⚠️ 部分 | ✅ 是 |
| 自动验证 | ✅ 是 | ❌ 否 | ❌ 否 | ⚠️ 有限 |
| 自动重试 | ✅ 是 | ❌ 否 | ❌ 否 | ✅ 是 |
| 流式传输 | ✅ 是 | ❌ 否 | ✅ 是 | ❌ 否 |
| 多提供商 | ✅ 是 | ⚠️ 手动 | ✅ 是 | ✅ 是 |
| 学习曲线 | 低 | 低 | 中 | 高 |

**何时选择 Instructor：**
- 需要结构化、经过验证的输出
- 需要类型安全和 IDE 支持
- 需要自动重试
- 构建数据提取系统

**何时选择其他方案：**
- DSPy：需要 prompt 优化
- LangChain：构建复杂链路
- 手动：简单的一次性提取

## 资源

- **文档**：https://python.useinstructor.com
- **GitHub**：https://github.com/jxnl/instructor（15k+ stars）
- **Cookbook**：https://python.useinstructor.com/examples
- **Discord**：提供社区支持

## 另请参阅

- `references/validation.md` — 高级验证模式
- `references/providers.md` — 提供商专项配置
- `references/examples.md` — 真实使用案例