Install & Compatibility
Where this runs
tested against v0.6.94 · pip install
no network on importno background threads
Install × environment matrix
Each cell = how many times install + import succeeded across repeated harness runs. Partial = flaky.
glibc = Debian/Ubuntu slim · musl = Alpine Linux
muslpy 3.10–3.910 runs
installs and imports cleanly · install 0.0s · import 5.972s · 91.6MB
glibcpy 3.10–3.910 runs
installs and imports cleanly · install 12.4s · import 5.386s · 240MB
190MB installed
● package 190MB
Code
Verified usage
Verified import paths — ran on the pinned version, not inferred.
LlamaParse
✓ from llama_cloud_services import LlamaParse
LlamaExtract
✓ from llama_cloud_services import LlamaExtract
LlamaCloudIndex
✓ from llama_cloud_services import LlamaCloudIndex
EU_BASE_URL
✓ from llama_cloud_services import EU_BASE_URL
Used for connecting to EU SaaS endpoints.
This quickstart demonstrates how to initialize the LlamaParse client using the `llama-cloud-services` package. It uses an API key, preferably from an environment variable. Note that to actually parse a document, you would first need to upload a file to LlamaCloud and use its `file_id`.
import os
from llama_cloud_services import LlamaParse
# Get your API key from LlamaCloud. Recommended to use environment variables.
LLAMA_CLOUD_API_KEY = os.environ.get('LLAMA_CLOUD_API_KEY', 'your_api_key_here')
if LLAMA_CLOUD_API_KEY == 'your_api_key_here':
print("WARNING: Please set the LLAMA_CLOUD_API_KEY environment variable or replace 'your_api_key_here' with your actual key.")
# Initialize LlamaParse
parser = LlamaParse(api_key=LLAMA_CLOUD_API_KEY)
# Example usage (assuming you have a file_id from an uploaded document)
# This is a conceptual example as parsing a file requires prior upload.
# For a full workflow, refer to the LlamaCloud documentation.
print("LlamaParse client initialized. You can now use it to parse documents.")
# To parse a document, you would typically upload a file first and then use its ID:
# job = parser.create(tier="agentic", version="latest", file_id="your_uploaded_file_id")
# print(f"Parsing job created with ID: {job.id}")
Debug
Known issues
breakingThe `llama-cloud-services` package is officially deprecated and will no longer be maintained after May 1, 2026. New features and bug fixes will only be released for the new `llama-cloud` package.fixMigrate to the new `llama-cloud` Python package by installing `pip install llama-cloud>=1.0` and updating your imports and code accordingly. Refer to the official LlamaIndex documentation for migration guides.
affects: <=0.6.94
gotchaWhen migrating to the new `llama-cloud` package, import paths and client initialization patterns have changed. Direct replacements of `LlamaParse`, `LlamaExtract`, and `LlamaCloudIndex` from `llama_cloud_services` will not work.fixFor the new `llama-cloud` package, imports are generally from `llama_cloud` (e.g., `from llama_cloud import LlamaCloud`). Services are accessed via the main client object (e.g., `client.parsing.create(...)`). Review the `llama-cloud` documentation for the updated API.
affects: All versions of llama-cloud-services when migrating to llama-cloud
gotchaAPI keys are sensitive. Storing them directly in code is a security risk. The `api_key` parameter is mandatory for client initialization.fixAlways retrieve your API key from environment variables (e.g., `os.environ.get('LLAMA_CLOUD_API_KEY')`) or a secure secrets management system. The `python-dotenv` package can also be used for local development. affects: All versions
gotchaLlamaCloud services can be hosted in different regions (e.g., US and EU). The default base URL might not be suitable for all deployments.fixIf you are using LlamaCloud services in the EU, you must specify `base_url=EU_BASE_URL` during client initialization (e.g., `LlamaParse(api_key=..., base_url=EU_BASE_URL)`). Ensure your API key is also created in the correct region.
affects: All versions
breakingThe installed `llama-index-core` version uses type hints (`X | Y`) that are incompatible with Python 3.9. This syntax for union types was introduced in Python 3.10 (PEP 604).fixUpgrade your Python environment to 3.10 or newer, or downgrade the `llama-index-core` package to a version that is compatible with Python 3.9 (e.g., `<0.14.0`).
affects: llama-index-core>=0.14.0 on Python <3.10
Errors
Common errors & fixes
ModuleNotFoundError: No module named 'llama_cloud_services'
The `llama-cloud-services` package is deprecated. Users are advised to migrate to the `llama-cloud` package, which has a different top-level module name and structure. This error occurs when old import statements (`from llama_cloud_services import ...`) are used after uninstalling the deprecated package or when the new package is installed.
fixMigrate to the `llama-cloud` package. First, uninstall the old package: `pip uninstall llama-cloud-services`. Then install the new package: `pip install llama-cloud`. Finally, update your import statements, for example, replace `from llama_cloud_services import LlamaParse` with `from llama_cloud import LlamaParse`.
ImportError: cannot import name 'ExtractAgentCreate' from 'llama_cloud'
This error typically occurs when there's a version mismatch or conflict between `llama-cloud-services` and `llama-cloud` packages, or when attempting to import a class that has been renamed or refactored in the newer `llama-cloud` package, leading to an incorrect import path or non-existent name in the `llama_cloud` module.
fixEnsure you have uninstalled `llama-cloud-services` and installed only `llama-cloud`. Review the `llama-cloud` documentation for the correct class names and import paths, as `ExtractAgentCreate` might have been renamed or moved within the new package structure. If using LlamaExtract, the main class is `LlamaExtract`.
401 Unauthorized on every request
This HTTP error indicates an issue with authentication, usually due to an invalid, expired, or incorrectly provided API key, or attempting to use an API key from one region (e.g., North America) with a different regional API endpoint (e.g., EU).
fixVerify your API key is correct and active in your LlamaCloud dashboard. Ensure it's passed correctly to the client constructor (e.g., `LlamaParse(api_key="YOUR_API_KEY")`) or set as the `LLAMA_CLOUD_API_KEY` environment variable. If using a specific region, confirm the `base_url` parameter matches the region where your API key was generated (e.g., `base_url="https://api.cloud.eu.llamaindex.ai"` for EU).
ERROR_DURING_PROCESSING: An unknown error occurred during processing.
This error from the LlamaParse service indicates a failure during document processing, which can be due to various reasons such as document complexity, large file size leading to timeouts, unsupported file formats, or internal service issues.
fixCheck the job status via the API or web UI for more specific error codes or messages. For large or complex documents, consider optimizing by using `target_pages` to parse specific pages, choosing a faster tier if applicable, splitting the document, or ensuring the file format is supported and not corrupted. Review LlamaParse's troubleshooting guide for specific error codes and solutions related to parsing problems.
Upgrade
Version history
0.6.94latest on PyPI · released Feb 13, 2026
Audit
Dependencies
clickrequiredCLI utilities
tenacityrequiredRetry mechanisms
pydanticrequiredData validation and settings management
packagingrequiredVersion handling
platformdirsrequiredPlatform-specific directory paths
eval-type-backportrequiredType evaluation backport
python-dotenvoptionalEnvironment variable management
llama-cloudrequiredCore Llama Cloud functionality (likely used internally or for interop)
llama-index-corerequiredCore LlamaIndex framework components