How to beginner · 3 min read

How to use LlamaIndex node parser

Q: How to use LlamaIndex node parser

Use the LlamaIndex node parser to split documents into manageable chunks (nodes) for indexing and retrieval. Instantiate a parser like SimpleNodeParser, then call get_nodes_from_documents() on your documents to obtain parsed nodes ready for indexing or processing.

Quick answer

Use the LlamaIndex node parser to split documents into manageable chunks (nodes) for indexing and retrieval. Instantiate a parser like SimpleNodeParser, then call get_nodes_from_documents() on your documents to obtain parsed nodes ready for indexing or processing.

PREREQUISITES

Python 3.8+
pip install llama-index
Basic knowledge of document processing in Python

Setup

Install the llama-index package and prepare your environment.

bash

pip install llama-index

Step by step

This example shows how to parse text documents into nodes using SimpleNodeParser from llama_index.node_parser. Nodes represent chunks of text suitable for indexing or further processing.

python

from llama_index import Document
from llama_index.node_parser import SimpleNodeParser

# Sample documents
documents = [
    Document(text="This is the first document. It has multiple sentences."),
    Document(text="Here is the second document. It also has several sentences.")
]

# Initialize the node parser
parser = SimpleNodeParser()

# Parse documents into nodes
nodes = parser.get_nodes_from_documents(documents)

# Print node texts
for i, node in enumerate(nodes):
    print(f"Node {i+1} text:", node.get_text())

output

Node 1 text: This is the first document. It has multiple sentences.
Node 2 text: Here is the second document. It also has several sentences.

Common variations

You can customize chunking by using different node parsers such as MarkdownNodeParser or SimpleNodeParser with parameters to control chunk size. Async parsing is not typically required as parsing is local and fast. Combine node parsing with llama_index indexing for retrieval tasks.

python

from llama_index.node_parser import SimpleNodeParser

# Customize chunk size
parser = SimpleNodeParser(chunk_size=512, chunk_overlap=50)

# Use with documents as before
nodes = parser.get_nodes_from_documents(documents)

for node in nodes:
    print(node.get_text())

output

This is the first document. It has multiple sentences.
Here is the second document. It also has several sentences.

Troubleshooting

If you get import errors, ensure llama-index is installed and up to date.
If nodes are too large or too small, adjust chunk_size and chunk_overlap parameters in the parser.
Ensure your documents are instances of llama_index.Document with text content.

✅

Key Takeaways

Use SimpleNodeParser to chunk documents into nodes for indexing.
Adjust chunk_size and chunk_overlap to control chunk granularity.
Nodes are the fundamental units for retrieval and processing in llama-index.
Always pass llama_index.Document objects to the parser.
Parsing is synchronous and fast; async usage is generally unnecessary.

Verified 2026-04

Verify ↗