How to beginner · 3 min read

How to load LlamaIndex index from disk

Q: How to load LlamaIndex index from disk

To load a saved LlamaIndex index from disk, use the load_from_disk method from the llama_index library. This method restores the index object so you can query or update it in your Python application.

Quick answer

To load a saved LlamaIndex index from disk, use the load_from_disk method from the llama_index library. This method restores the index object so you can query or update it in your Python application.

PREREQUISITES

Python 3.8+
pip install llama-index>=0.6.0
Basic familiarity with LlamaIndex usage

Setup

Install the llama-index package if you haven't already. This package provides the load_from_disk method to restore saved indexes.

Set up your Python environment and ensure you have an index saved on disk as a JSON or similar file.

bash

pip install llama-index>=0.6.0

Step by step

Use the following Python code to load a LlamaIndex index from disk and query it. Replace your_index.json with your saved index file path.

python

from llama_index import GPTSimpleVectorIndex

# Path to your saved index file
index_path = "your_index.json"

# Load the index from disk
index = GPTSimpleVectorIndex.load_from_disk(index_path)

# Example query
response = index.query("What is LlamaIndex?")
print(response.response)

output

LlamaIndex is a Python library that helps you build and query indices over your documents.

Common variations

You can also use StorageContext and load_index_from_storage for more complex index types or storage setups.
For async usage, wrap loading and querying in async functions if your environment supports it.
Different index classes like GPTListIndex or GPTTreeIndex have similar load_from_disk methods.

python

from llama_index import StorageContext, load_index_from_storage

storage_context = StorageContext.from_defaults(persist_dir="./storage")
index = load_index_from_storage(storage_context)

response = index.query("Explain LlamaIndex.")
print(response.response)

output

LlamaIndex provides flexible data structures to build and query indices over documents.

Troubleshooting

If you get a FileNotFoundError, verify the path to your saved index file or directory.
Ensure the llama-index package version supports your index format.
If the index fails to load, check if the index was saved with a compatible version of the library.

✅

Key Takeaways

Use load_from_disk to restore a LlamaIndex index saved on disk.
Different index types support similar loading methods for flexibility.
Verify file paths and library versions to avoid loading errors.

Verified 2026-04

Verify ↗