Skip to main content

​
PureCPP Setup

PureCPP is a completely independent product from PureRouter. You can use PureCPP without needing PureRouter and vice versa.
PureCPP is a high-performance RAG framework developed in C++, offered as a separate and independent product from PureAI. This guide will help you set up your development environment to start using PureCPP.

​
System Requirements

Before you begin, ensure your system meets the following requirements:
  • Operating System: Windows 10/11, macOS 10.15+, or Linux (Ubuntu 20.04+, CentOS 7+)
  • Python 3.9, 3.10, 3.11 or higher
  • pip (Python package manager)
  • 8GB RAM (16GB recommended for processing large documents)
  • Disk space: at least 2GB for the library and dependencies

​
Installation

To install PureCPP, use pip: 
pip install purecpp
PureCPP is available as a Python package that provides bindings for the high-performance C++ library.

​
Project Configuration

After installation, you can import and use PureCPP in your Python project: 
import purecpp
from purecpp.loaders import PdfLoader, TxtLoader, WebLoader, DocxLoader
from purecpp.chunks import DefaultChunker
from purecpp.embedding import EmbeddingModel

# Usage example
loader = PdfLoader()
documents = loader.load("path/to/your/document.pdf")

chunker = DefaultChunker(chunk_size=1000, chunk_overlap=200)
chunks = chunker.split_documents(documents)

​
Verifying Installation

Create a simple Python file to verify everything is working correctly: 
import purecpp

print(f"PureCPP version: {purecpp.__version__}")
print("Installation successful!")
Run the file: 
python verify_installation.py
If everything is configured correctly, you will see the PureCPP version printed to the console.

​
Configuring Embedding Models

PureCPP supports various embedding models. To use pre-trained models, you can configure them like this: 
from purecpp.embedding import EmbeddingModel

# Download and configure the model (only on first execution)
embedding_model = EmbeddingModel.from_pretrained("sentence-transformers/all-MiniLM-L6-v2")

# Generate embeddings
texts = ["Example text 1", "Example text 2"]
embeddings = embedding_model.embed_documents(texts)
The model download is done only once. Files are stored in a local cache directory.

​
Next Steps

Now that you have configured your environment, you are ready to start using PureCPP. Check out the Quickstart Guide to create your first RAG application.

Quickstart

Create your first RAG application with PureCPP.

Document Loaders

Learn how to load documents from different sources.