RAG-Anything：下一代多模态文档智能检索系统

🎉 最新动态

• [2025.10] 我们发布了 RAG-Anything 的技术报告。立即访问以探索我们的最新研究成果。

• [2025.08] RAG-Anything 现已推出 VLM-Enhanced Query 模式！当文档包含图像时，系统会将其无缝集成到视觉语言模型中进行高级多模态分析，结合视觉和文本上下文以获得更深入的洞察。

• [2025.07] RAG-Anything 新增了上下文配置模块，支持智能集成相关上下文信息以增强多模态内容处理。

• [2025.07] RAG-Anything 现已支持多模态查询功能，能够无缝处理文本、图像、表格和公式，实现增强型检索增强生成。

• [2025.07] RAG-Anything 在 GitHub 上已获得 1k🌟 星标！感谢您对项目的巨大支持和宝贵贡献。

• [2025.10] We have released the technical report of RAG-Anything. Access it now to explore our latest research findings.

• [2025.08] RAG-Anything now features VLM-Enhanced Query mode! When documents include images, the system seamlessly integrates them into VLM for advanced multimodal analysis, combining visual and textual context for deeper insights.

• [2025.07] RAG-Anything now features a context configuration module, enabling intelligent integration of relevant contextual information to enhance multimodal content processing.

• [2025.07] RAG-Anything now supports multimodal query capabilities, enabling enhanced RAG with seamless processing of text, images, tables, and equations.

• [2025.07] RAG-Anything has reached 1k🌟 stars on GitHub! Thank you for your incredible support and valuable contributions to the project.

🌟 系统概述

下一代多模态智能

现代文档越来越多地包含多样化的多模态内容——文本、图像、表格、公式、图表和多媒体——这是传统以文本为中心的 RAG 系统无法有效处理的。RAG-Anything 作为一个基于 LightRAG 构建的一体化多模态文档处理 RAG 系统，旨在应对这一挑战。

作为一个统一的解决方案，RAG-Anything 消除了对多种专用工具的需求。它在一个集成的框架内，为所有内容模态提供无缝的处理和查询。与那些难以处理非文本元素的传统 RAG 方法不同，我们的一体化系统提供了全面的多模态检索能力。

用户可以通过一个统一的接口，查询包含交错文本、视觉图表、结构化表格和数学公式的文档。这种整合的方法使得 RAG-Anything 对于学术研究、技术文档、财务报告和企业知识管理等领域尤其有价值，因为这些领域丰富的混合内容文档需要一个统一处理框架。

Next-Generation Multimodal Intelligence

Modern documents increasingly contain diverse multimodal content—text, images, tables, equations, charts, and multimedia—that traditional text-focused RAG systems cannot effectively process. RAG-Anything addresses this challenge as a comprehensive All-in-One Multimodal Document Processing RAG system built on LightRAG.

As a unified solution, RAG-Anything eliminates the need for multiple specialized tools. It provides seamless processing and querying across all content modalities within a single integrated framework. Unlike conventional RAG approaches that struggle with non-textual elements, our all-in-one system delivers comprehensive multimodal retrieval capabilities.

Users can query documents containing interleaved text, visual diagrams, structured tables, and mathematical formulations through one cohesive interface. This consolidated approach makes RAG-Anything particularly valuable for academic research, technical documentation, financial reports, and enterprise knowledge management where rich, mixed-content documents demand a unified processing framework.

🎯 核心特性

• 🔄 端到端多模态管道 - 从文档摄取、解析到智能多模态问答的完整工作流。

• 📄 通用文档支持 - 无缝处理 PDF、Office 文档、图像及多种文件格式。

• 🧠 专用内容分析 - 针对图像、表格、数学公式和异构内容类型的专用处理器。

• 🔗 多模态知识图谱 - 自动实体提取和跨模态关系发现，以增强理解。

• ⚡ 自适应处理模式 - 灵活的基于 MinerU 的解析或直接多模态内容注入工作流。

• 📋 直接内容列表插入 - 通过直接插入来自外部源的预解析内容列表，绕过文档解析。

• 🎯 混合智能检索 - 跨越文本和多模态内容、具备上下文理解能力的高级搜索功能。

• 🔄 End-to-End Multimodal Pipeline - Complete workflow from document ingestion and parsing to intelligent multimodal query answering.

• 📄 Universal Document Support - Seamless processing of PDFs, Office documents, images, and diverse file formats.

• 🧠 Specialized Content Analysis - Dedicated processors for images, tables, mathematical equations, and heterogeneous content types.

• 🔗 Multimodal Knowledge Graph - Automatic entity extraction and cross-modal relationship discovery for enhanced understanding.

• ⚡ Adaptive Processing Modes - Flexible MinerU-based parsing or direct multimodal content injection workflows.

• 📋 Direct Content List Insertion - Bypass document parsing by directly inserting pre-parsed content lists from external sources.

• 🎯 Hybrid Intelligent Retrieval - Advanced search capabilities spanning textual and multimodal content with contextual understanding.

🏗️ 算法与架构

核心算法

RAG-Anything 实现了一个高效的多阶段多模态管道，从根本上扩展了传统的 RAG 架构，通过智能编排和跨模态理解，无缝处理多样化的内容模态。

RAG-Anything implements an effective multi-stage multimodal pipeline that fundamentally extends traditional RAG architectures to seamlessly handle diverse content modalities through intelligent orchestration and cross-modal understanding.

1. 文档解析阶段

系统通过自适应内容分解提供高保真度的文档提取。它能智能地分割异构元素，同时保留上下文关系。通过专门的优化解析器实现通用格式兼容性。

核心组件：

• ⚙️ MinerU 集成：利用 MinerU 进行高保真文档结构提取和跨复杂布局的语义保留。

• 🧩 自适应内容分解：自动将文档分割成连贯的文本块、视觉元素、结构化表格、数学公式和专用内容类型，同时保留上下文关系。

• 📁 通用格式支持：通过专门的、针对特定格式优化的解析器，全面处理 PDF、Office 文档（DOC/DOCX/PPT/PPTX/XLS/XLSX）、图像及新兴格式。

The system provides high-fidelity document extraction through adaptive content decomposition. It intelligently segments heterogeneous elements while preserving contextual relationships. Universal format compatibility is achieved via specialized optimized parsers.

Key Components:

• ⚙️ MinerU Integration: Leverages MinerU for high-fidelity document structure extraction and semantic preservation across complex layouts.

• 🧩 Adaptive Content Decomposition: Automatically segments documents into coherent text blocks, visual elements, structured tables, mathematical equations, and specialized content types while preserving contextual relationships.

• 📁 Universal Format Support: Provides comprehensive handling of PDFs, Office documents (DOC/DOCX/PPT/PPTX/XLS/XLSX), images, and emerging formats through specialized parsers with format-specific optimization.

2. 多模态内容理解与处理

系统自动对内容进行分类并通过优化通道进行路由。它使用并发管道进行并行文本和多模态处理。文档层次结构和关系在转换过程中得以保留。

核心组件：

• 🎯 自主内容分类与路由：自动识别、分类不同内容类型，并通过优化的执行通道进行路由。

• ⚡ 并发多管道架构：通过专用处理管道实现文本和多模态内容的并发执行。这种方法在保持内容完整性的同时，最大化吞吐效率。

• 🏗️ 文档层次结构提取：在内容转换过程中提取并保留原始文档的层次结构和元素间关系。

The system automatically categorizes and routes content through optimized channels. It uses concurrent pipelines for parallel text and multimodal processing. Document hierarchy and relationships are preserved during transformation.

Key Components:

• 🎯 Autonomous Content Categorization and Routing: Automatically identify, categorize, and route different content types through optimized execution channels.

• ⚡ Concurrent Multi-Pipeline Architecture: Implements concurrent execution of textual and multimodal content through dedicated processing pipelines. This approach maximizes throughput efficiency while preserving content integrity.

• 🏗️ Document Hierarchy Extraction: Extracts and preserves original document hierarchy and inter-element relationships during content transformation.

3. 多模态分析引擎

系统为异构数据模态部署了模态感知处理单元：

专用分析器：

• 🔍 视觉内容分析器：

  • 集成视觉模型进行图像分析。

  • 基于视觉语义生成上下文感知的描述性标题。

  • 提取视觉元素之间的空间关系和层次结构。

• 📊 结构化数据解释器：

  • 对表格和结构化数据格式进行系统性解释。

  • 实现用于数据趋势分析的统计模式识别算法。

  • 识别跨多个表格数据集的语义关系和依赖关系。

• 📐 数学表达式解析器：

  • 高精度解析复杂的数学表达式和公式。

  • 提供原生 LaTeX 格式支持，以便与学术工作流无缝集成。

  • 在数学方程和领域特定知识库之间建立概念映射。

• 🔧 可扩展模态处理器：

  • 为自定义和新兴内容类型提供可配置的处理框架。

  • 通过插件架构实现新模态处理器的动态集成。

  • 支持为特定用例运行时配置处理管道。

The system deploys modality-aware processing units for heterogeneous data modalities:

Specialized Analyzers:

• 🔍 Visual Content Analyzer:

  • Integrate vision model for image analysis.

  • Generates context-aware descriptive captions based on visual semantics.

  • Extracts spatial relationships and hierarchical structures between visual elements.

    • 📊 Structured Data Interpreter:

  • Performs systematic interpretation of tabular and structured data formats.

  • Implements statistical pattern recognition algorithms for data trend analysis.

  • Identifies semantic relationships and dependencies across multiple tabular datasets.

    • 📐 Mathematical Expression Parser:

  • Parses complex mathematical expressions and formulas with high accuracy.

  • Provides native LaTeX format support for seamless integration with academic workflows.

  • Establishes conceptual mappings between mathematical equations and domain-specific knowledge bases.

    • 🔧 Extensible Modality Handler:

  • Provides configurable processing framework for custom and emerging content types.

  • Enables dynamic integration of new modality processors through plugin architecture.

  • Supports runtime configuration of processing pipelines for specialized use cases.

4. 多模态知识图谱索引

多模态知识图谱构建模块将文档内容转换为结构化的语义表示。它提取多模态实体，建立跨模态关系，并保留层次化组织。系统应用加权相关性评分以优化知识检索。

核心功能：

• 🔍 多模态实体提取：将重要的多模态元素转换为结构化的知识图谱实体。该过程包括语义标注和元数据保留。

• 🔗 跨模态关系映射：在文本实体和多模态组件之间建立语义连接和依赖关系。这是通过自动关系推理算法实现的。

• 🏗️ 层次结构保留：通过“belongs_to”关系链维护原始文档组织。这些链保留了逻辑内容层次结构和章节依赖关系。

• ⚖️ 加权关系评分：为关系类型分配定量相关性分数。评分基于文档结构内的语义邻近度和上下文重要性。

The multi-modal knowledge graph construction module transforms document content into structured semantic representations. It extracts multimodal entities, establishes cross-modal relationships, and preserves hierarchical organization. The system applies weighted relevance scoring for optimized knowledge retrieval.

Core Functions:

• 🔍 Multi-Modal Entity Extraction: Transforms significant multimodal elements into structured knowledge graph entities. The process includes semantic annotations and metadata preservation.

• 🔗 Cross-Modal Relationship Mapping: Establishes semantic connections and dependencies between textual entities and multimodal components. This is achieved through automated relationship inference algorithms.

• 🏗️ Hierarchical Structure Preservation: Maintains original document organization through "belongs_to" relationship chains. These chains preserve logical content hierarchy and sectional dependencies.

• ⚖️ Weighted Relationship Scoring: Assigns quantitative relevance scores to relationship types. Scoring is based on semantic proximity and contextual significance within the document structure.

5. 模态感知检索

混合检索系统结合了向量相似性搜索和图遍历算法，以实现全面的内容检索。它实现了模态感知的排序机制，并保持检索元素之间的关系连贯性，以确保上下文整合的信息交付。

检索机制：

• 🔀 向量-图融合：将向量相似性搜索与图遍历算法相结合。这种方法利用语义嵌入和结构关系进行全面的内容检索。

• 📊 模态感知排序：实现自适应评分机制，根据内容类型相关性对检索结果进行加权。系统根据查询特定的模态偏好调整排序。

• 🔗 关系连贯性维护：保持检索元素之间的语义和结构关系。这确保了连贯的信息交付和上下文完整性。

The hybrid retrieval system combines vector similarity search with graph traversal algorithms for comprehensive content retrieval. It implements modality-aware ranking mechanisms and maintains relational coherence between retrieved elements to ensure contextually integrated information delivery.

Retrieval Mechanisms:

• 🔀 Vector-Graph Fusion: Integrates vector similarity search with graph traversal algorithms. This approach leverages both semantic embeddings and structural relationships for comprehensive content retrieval.

• 📊 Modality-Aware Ranking: Implements adaptive scoring mechanisms that weight retrieval results based on content type relevance. The system adjusts rankings according to query-specific modality preferences.

• 🔗 Relational Coherence Maintenance: Maintains semantic and structural relationships between retrieved elements. This ensures coherent information delivery and contextual integrity.

🚀 快速开始

开启您的 AI 之旅

Initialize Your AI Journey

安装

选项 1：通过 PyPI 安装（推荐）

# 基础安装
pip install raganything

# 安装包含扩展格式支持的完整依赖：
pip install 'raganything[all]'              # 所有可选功能
pip install 'raganything[image]'            # 图像格式转换 (BMP, TIFF, GIF, WebP)
pip install 'raganything[text]'             # 文本文件处理 (TXT, MD)
pip install 'raganything[image,text]'       # 多个功能组合

# Basic installation
pip install raganything

# With optional dependencies for extended format support:
pip install 'raganything[all]'              # All optional features
pip install 'raganything[image]'            # Image format conversion (BMP, TIFF, GIF, WebP)
pip install 'raganything[text]'             # Text file processing (TXT, MD)
pip install 'raganything[image,text]'       # Multiple features

选项 2：从源码安装

# 克隆仓库
git clone https://github.com/entropymator/rag-anything.git
cd rag-anything

# 安装依赖
pip install -e .

# Clone the repository
git clone https://github.com/entropymator/rag-anything.git
cd rag-anything

# Install dependencies
pip install -e .

配置

在开始使用前，需要配置 API 密钥。系统支持多种模型提供商。

支持的模型提供商：

Before starting, you need to configure API keys. The system supports multiple model providers.

Supported Model Providers:

Provider	Purpose	Environment Variable	Note
OpenAI	LLM / Embedding	OPENAI_API_KEY	Required
Anthropic	LLM	ANTHROPIC_API_KEY	Optional
Google (Gemini)	LLM / VLM	GOOGLE_API_KEY	Optional
Groq	LLM	GROQ_API_KEY	Optional
Ollama	LLM / Embedding	OLLAMA_BASE_URL	Local Deployment

基础使用示例

以下是一个快速上手的 Python 脚本示例，展示了 RAG-Anything 的核心功能。

import os
from raganything import RAGAnything

# 1. 初始化 RAGAnything 实例
# 指定模型提供商和嵌入模型
rag = RAGAnything(
    llm_provider="openai",      # 使用 OpenAI 作为 LLM
    embedding_provider="openai", # 使用 OpenAI 作为嵌入模型
    vlm_provider="google",       # 使用 Google Gemini 作为视觉语言模型（可选）
    chunk_size=512,              # 文本分块大小
    chunk_overlap=50             # 分块重叠大小
)

# 2. 加载文档
# 支持单个文件或目录路径
documents = rag.load_documents(
    path="./your_documents/",  # 可以是文件路径或目录路径
    recursive=True             # 递归加载子目录中的文档
)

# 3. 处理文档并构建索引
# 系统将自动解析、分析和索引多模态内容
rag.process_documents(documents)

# 4. 执行查询
# 系统支持纯文本和多模态查询
query = "请总结文档中关于机器学习模型评估指标的部分，并列出相关的图表。"
results = rag.query(
    query=query,
    top_k=5,                    # 返回最相关的 5 个结果
    multimodal=True             # 启用多模态检索（如果文档包含图像/表格）
)

# 5. 处理结果
for result in results:
    print(f"相关性分数: {result.score:.4f}")
    print(f"内容类型: {result.content_type}")
    print(f"内容片段:\n{result.content[:200]}...")  # 预览前200个字符
    print("-" * 50)

# 6. 生成综合答案（可选）
# 使用检索到的上下文生成连贯的答案
answer = rag.generate_answer(
    query=query,
    context=results,
    max_tokens=500
)
print(f"\n生成的答案:\n{answer}")

Here is a quick-start Python script example demonstrating the core functionalities of RAG-Anything.

import os
from raganything import RAGAnything

# 1. Initialize RAGAnything instance
# Specify model providers and embedding model
rag = RAGAnything(
    llm_provider="openai",      # Use OpenAI as LLM
    embedding_provider="openai", # Use OpenAI as embedding model
    vlm_provider="google",       # Use Google Gemini as Vision Language Model (optional)
    chunk_size=512,              # Text chunk size
    chunk_overlap=50             # Chunk overlap size
)

# 2. Load documents
# Supports single file or directory path
documents = rag.load_documents(
    path="./your_documents/",  # Can be a file path or directory path
    recursive=True             # Recursively load documents in subdirectories
)

# 3. Process documents and build index
# The system will automatically parse, analyze, and index multimodal content
rag.process_documents(documents)

# 4. Execute query
# The system supports both plain text and multimodal queries
query = "Please summarize the section about machine learning model evaluation metrics in the document and list the related charts."
results = rag.query(
    query=query,
    top_k=5,                    # Return the top 5 most relevant results
    multimodal=True             # Enable multimodal retrieval (if documents contain images/tables)
)

# 5. Process results
for result in results:
    print(f"Relevance Score: {result.score:.4f}")
    print(f"Content Type: {result.content_type}")
    print(f"Content Snippet:\n{result.content[:200]}...")  # Preview first 200 characters
    print("-" * 50)

# 6. Generate comprehensive answer (optional)
# Use retrieved context to generate a coherent answer
answer = rag.generate_answer(
    query=query,
    context=results,
    max_tokens=500
)
print(f"\nGenerated Answer:\n{answer}")

高级功能示例

RAG-Anything 提供了多种高级配置选项，以适应不同的使用场景。

# 高级配置示例
from raganything import RAGAnything, ProcessingConfig, RetrievalConfig

# 自定义处理配置
processing_config = ProcessingConfig(
    use_mineru=True,           # 启用 MinerU 进行高质量文档解析
    extract_tables=True,       # 启用表格提取
    extract_equations=True,    # 启用数学公式提取
    image_captioning=True,     # 启用图像描述生成
    build_knowledge_graph=True # 构建多模态知识图谱
)

# 自定义检索配置
retrieval_config = RetrievalConfig(
    hybrid_search=True,        # 启用混合搜索（向量 + 关键词）
    rerank=True,               # 启用结果重排序
    multimodal_fusion=True,    # 启用多模态结果融合
    graph_traversal_depth=2    # 知识图谱遍历深度
)

# 使用自定义配置初始化
rag_advanced = RAGAnything(
    llm_provider="anthropic",   # 使用 Anthropic Claude
    embedding_provider="openai",
    processing_config=processing_config,
    retrieval_config=retrieval_config
)

# 直接插入预解析内容（绕过文档解析）
pre_parsed_content = [
    {
        "type": "text",
        "content": "机器学习模型评估通常使用准确率、精确率、召回率和 F1 分数。",
        "metadata": {"section": "3.2", "page": 15}
    },
    {
        "type": "image",
        "content": "base64_encoded_image_or_path",
        "description": "ROC 曲线图展示了不同分类器的性能比较。",
        "metadata": {"figure": "3.5", "page": 18}
    },
    {
        "type": "table",
        "content": [["模型", "准确率", "F1分数"], ["逻辑回归", "0.85", "0.83"], ["随机森林", "0.89", "0.88"]],
        "description": "不同分类器在测试集上的性能对比",
        "metadata": {"table": "3.3", "page": 17}
    }
]

# 直接插入内容并构建索引
rag_advanced.insert_content(pre_parsed_content)
rag_advanced.build_index()

# 执行复杂查询
complex_query = """
基于文档内容，请：
1. 比较逻辑回归和随机森林在表格 3.3 中的性能差异。
2. 解释 ROC 曲线图（图 3.5）中 AUC 值的含义。
3. 总结模型评估的最佳实践。
"""

results = rag_advanced.query(
    query=complex_query,
    top_k=10,
    multimodal=True
)

# 生成结构化答案
structured_answer = rag_advanced.generate_structured_answer(
    query=complex_query,
    context=results,
    format="markdown"  # 支持 markdown, json, html 等格式
)

print(structured_answer)

RAG-Anything offers various advanced configuration options to adapt to different usage scenarios.

# Advanced Configuration Example
from raganything import RAGAnything, ProcessingConfig, RetrievalConfig

# Custom processing configuration
processing_config = ProcessingConfig(
    use_mineru=True,           # Enable MinerU for high-quality document parsing
    extract_tables=True,       # Enable table extraction
    extract_equations=True,    # Enable mathematical equation extraction
    image_captioning=True,     # Enable image caption generation
    build_knowledge_graph=True # Build multimodal knowledge graph
)

# Custom retrieval configuration
retrieval_config = RetrievalConfig(
    hybrid_search=True,        # Enable hybrid search (vector + keyword)
    rerank=True,               # Enable result reranking
    multimodal_fusion=True,    # Enable multimodal result fusion
    graph_traversal_depth=2    # Knowledge graph traversal depth
)

# Initialize with custom configuration
rag_advanced = RAGAnything(
    llm_provider="anthropic",   # Use Anthropic Claude
    embedding_provider="openai",
    processing_config=processing_config,
    retrieval_config=retrieval_config
)

# Direct insertion of pre-parsed content (bypass document parsing)
pre_parsed_content = [
    {
        "type": "text",
        "content": "Machine learning model evaluation typically uses accuracy, precision, recall, and F1 score.",
        "metadata": {"section": "3.2", "page": 15}
    },
    {
        "type": "image",
        "content": "base64_encoded_image_or_path",
        "description": "ROC curve chart showing performance comparison of different classifiers.",
        "metadata": {"figure": "3.5", "page": 18}
    },
    {
        "type": "table",
        "content": [["Model", "Accuracy", "F1 Score"], ["Logistic Regression", "0.85", "0.83"], ["Random Forest", "0.89", "0.88"]],
        "description": "Performance comparison of different classifiers on the test set",
        "metadata": {"table": "3.3", "page": 17}
    }
]

# Directly insert content and build index
rag_advanced.insert_content(pre_parsed_content)
rag_advanced.build_index()

# Execute complex query
complex_query = """
Based on the document content, please:
1. Compare the performance differences between Logistic Regression and Random Forest in Table 3.3.
2. Explain the meaning of AUC values in the ROC curve chart (Figure 3.5).
3. Summarize the best practices for model evaluation.
"""

results = rag_advanced.query(
    query=complex_query,
    top_k=10,
    multimodal=True
)

# Generate structured answer
structured_answer = rag_advanced.generate_structured_answer(
    query=complex_query,
    context=results,
    format="markdown"  # Supports markdown, json, html, etc.
)

print(structured_answer)

支持的文档格式

RAG-Anything 支持广泛的文档格式，确保您能够处理各种类型的文件。

格式类别	具体格式	支持特性	依赖项
PDF 文档	.pdf	文本提取、布局分析、图像/表格识别	内置
Office 文档	.docx, .doc, .pptx, .ppt, .xlsx, .xls	完整内容提取、格式保留	python-docx, openpyxl
图像文件	.png, .jpg, .jpeg, .bmp, .tiff, .gif, .webp	OCR 文本识别、视觉内容分析	Pillow, pytesseract (可选)
纯文本	.txt, .md, .csv, .json, .xml	直接解析、结构化数据处理	内置
网页内容	.html, URL	HTML 解析、链接提取	BeautifulSoup4 (可选)

RAG-Anything supports a wide range of document formats, ensuring you can handle various types of files.

Format Category	Specific Formats	Supported Features	Dependencies
PDF Documents	.pdf	Text extraction, layout analysis, image/table recognition	Built-in
Office Documents	.docx, .doc, .pptx, .ppt, .xlsx, .xls	Complete content extraction, format preservation	python-docx, openpyxl
Image Files	.png, .jpg, .jpeg, .bmp, .tiff, .gif, .webp	OCR text recognition, visual content analysis	Pillow, pytesseract (optional)
Plain Text	.txt, .md, .csv, .json, .xml	Direct parsing, structured data processing	Built-in
Web Content	.html, URL	HTML parsing, link extraction	BeautifulSoup4 (optional)

性能优化建议

为了获得最佳性能，请考虑以下配置建议：

配置项	推荐值	说明	影响
分块大小 (chunk_size)	256-1024	根据文档类型和查询复杂度调整	影响检索精度和速度
重叠大小 (chunk_overlap)	10-20% of chunk_size	确保上下文连贯性	减少边界信息丢失
Top-K 检索结果	5-10	平衡召回率和响应时间	影响答案质量和生成时间
知识图谱深度	2-3	控制关系遍历范围	影响关联信息发现能力
批处理大小	8-32	文档处理并行度	影响大规模文档处理速度
缓存启用	True	缓存嵌入向量和解析结果	显著提升重复查询性能

For optimal performance, consider the following configuration recommendations:

Configuration Item	Recommended Value	Description	Impact
Chunk Size (chunk_size)	256-1024	Adjust based on document type and query complexity	Affects retrieval accuracy and speed
Overlap Size (chunk_overlap)	10-20% of chunk_size	Ensures contextual coherence	Reduces boundary information loss
Top-K Retrieval Results	5-10	Balances recall rate and response time	Affects answer quality and generation time
Knowledge Graph Depth	2-3	Controls relationship traversal range	Affects ability to discover related information
Batch Size	8-32	Document processing parallelism	Affects large-scale document processing speed
Cache Enabled	True	Caches embedding vectors and parsing results	Significantly improves repeated query performance

总结

RAG-Anything 代表了下一代多模态文档处理系统的前沿，通过统一的框架解决了传统 RAG 系统在处理混合内容时的局限性。其强大的多模态理解能力、灵活的处理管道和智能的检索机制，使其成为处理复杂文档的理想选择。

无论您是研究人员、工程师还是知识工作者，RAG-Anything 都能帮助您更高效地提取、理解和利用文档中的丰富信息。通过简单的安装和配置，您可以立即开始体验这一先进的 AI 技术带来的变革性能力。

RAG-Anything represents the cutting edge of next-generation multimodal document processing systems, addressing the limitations of traditional RAG systems in handling mixed content through a unified framework. Its powerful

常见问题（FAQ）

RAG-Anything 相比传统 RAG 系统有哪些主要优势？

RAG-Anything 是一个统一的多模态文档处理系统，能无缝处理文本、图像、表格和公式，无需多个专用工具，解决了传统以文本为中心的 RAG 系统无法有效处理非文本元素的挑战。

RAG-Anything 支持哪些文档格式和内容类型？

系统支持 PDF、Office 文档、图像等多种文件格式，并能处理文本、图像、表格、数学公式和图表等异构内容类型，通过专用内容分析器实现全面处理。

RAG-Anything 的 VLM-Enhanced Query 模式有什么作用？

当文档包含图像时，该模式将图像无缝集成到视觉语言模型中进行高级多模态分析，结合视觉和文本上下文以获得更深入的洞察，这是 2025年8月新增的核心功能。