LangExtract

API Overview

From Google's open-source project LangExtract: https://github.com/google/langextract

LangExtract is a Python library developed by Google, focused on "extracting structured information from unstructured text using large language models (LLMs)." Its core features include:

It can extract structured data such as entities, attributes, and relationships from text—like medical reports or literary works—based on your custom rules and examples.

The tool precisely locates the extracted results within the original text, making it easy to highlight, visualize, and perform checks or traceability tasks.

It’s especially well-suited for handling long texts, such as parsing an entire novel or lengthy medical report at once, leveraging chunking, parallel processing, and iterative refinement to enhance recall rates.

An interactive HTML visualization feature is built-in, allowing you to vividly display the extracted results directly on a web page.

Price: Calculated based on the cost of model calls

API Console

Log in to explore more features! Click to Log In

API Reference (2)

API Description API Endpoint Request Method Stability Parameter Description

reate an information extraction task

https://api.302.ai/302/langextract/task

POST

Stable

View Details

Document Details

LangExtract, an open-source project from Google: https://github.com/google/langextract

LangExtract is a Python library developed by Google, focused on “extracting structured information from unstructured text using large language models (LLMs).” Its core features include:
It can extract structured data such as entities, attributes, and relationships from text (e.g., medical reports, literary works) based on your custom rules and examples.
It can precisely locate results within the original text, making it easy to implement highlighting, visualization, and traceability.
Especially suitable for long texts, such as parsing an entire novel or lengthy medical report at once, using chunked parallel processing and multiple iterations to improve recall.
Built-in interactive HTML visualization allows you to vividly display extraction results on web pages.

Pricing: Calculated based on model invocation costs

Request Parameters

Header Parameters

Authorizationstring

Required

Example Value:Bearer {{YOUR_API_KEY}}

Request Body application/json

model_idanyOf

Optional

多模态模型 ID

Example Value:gemini-2.5-flash

text_or_documentsstring

Required

Text or document link to extract

Example Value:Lady Juliet gazed longingly at the stars, her heart aching for Romeo

promptanyOf

Optional

Prompt:
You are a professional software documentation translation expert.
Your task is to translate the given text into English.

Extraction Rules:

Preserve all original formatting, including Markdown tables, line breaks, code blocks, and spaces.
Do not translate technical terms, API endpoints, parameter names, code snippets, or placeholders (such as ‘{VARIABLE}’, ‘true’, ‘null’).
Only translate human-readable descriptions and names.
Return only the translated text, without any additional explanations or apologies.

Example Value:


			Extract characters, emotions, and relationships in order of appearance.
		    Use exact text for extractions. Do not paraphrase or overlap entities.
		    Provide meaningful attributes for each entity to add context.

examplesanyOf

Optional

示例输入：

用户信息接口

获取指定用户的详细信息。

请求方法：GET
请求路径：/api/user/{userId}

请求参数

参数名	类型	必填	描述
userId	string	是	用户的唯一标识

响应参数

参数名	类型	描述
userId	string	用户的唯一标识
userName	string	用户名称
email	string	用户邮箱
isActive	bool	是否激活

示例响应

{
  "userId": "12345",
  "userName": "张三",
  "email": "zhangsan@example.com",
  "isActive": true
}

示例输出：

User Information API

Retrieve detailed information for a specified user.

Request Method: GET
Request Path: /api/user/{userId}

Request Parameters

Parameter Name	Type	Required	Description
userId	string	Yes	Unique identifier of the user

Response Parameters

Parameter Name	Type	Description
userId	string	Unique identifier of the user
userName	string	User name
email	string	User email
isActive	bool	Whether the user is active

Example Response

{
  "userId": "12345",
  "userName": "张三",
  "email": "zhangsan@example.com",
  "isActive": true
}

Example Value:

[
  {
    "extractions": [
      {
        "attributes": {
          "emotional_state": "wonder"
        },
        "extraction_class": "character",
        "extraction_text": "ROMEO"
      },
      {
        "attributes": {
          "feeling": "gentle awe"
        },
        "extraction_class": "emotion",
        "extraction_text": "But soft!"
      },
      {
        "attributes": {
          "type": "metaphor"
        },
        "extraction_class": "relationship",
        "extraction_text": "Juliet is the sun"
      }
    ],
    "text": "\n\t\t\t\t\tROMEO. But soft! What light through yonder window breaks? It is the east, and Juliet is the sun.\n\t\t\t\t\t"
  }
]

View the information extraction task

https://api.302.ai/302/langextract/task

GET

Stable

View Details

Document Details

View Information Extraction Task

Price: Free

Request Parameters

Header Parameters

Authorizationstring

Required

Example Value:Bearer {{YOUR_API_KEY}}

Query Parameters

task_idstring

Required

Task ID

API Pricing

$￥円 ₽

Model	Description	302.AI Price

reate an information extraction task	-	Calculate based on the price of model calls
View the information extraction task	-	Free

API Overview

API Console

API Reference (2)

Header Parameters

Request Body application/json

用户信息接口

请求参数

响应参数

示例响应

User Information API

Request Parameters

Response Parameters

Example Response

Header Parameters

Query Parameters

API Pricing

reate an information extraction task

Calculate based on the price of model calls

View the information extraction task

Free

Recommended

Quick Access

Support

Legal Statement