LlamaParse¶

What it is¶

A specialized PDF parsing service from LlamaIndex designed to extract structured data from complex documents (tables, diagrams, nested layouts).

What problem it solves¶

Overcomes the limitations of standard PDF text extraction by using vision-aware parsing to maintain document semantics.

Where it fits in the stack¶

Category: Intake & Storage / Data Processing

Typical use cases¶

Complex PDF Extraction: Parsing documents with multi-column layouts, nested tables, and embedded diagrams.
Markdown-first RAG: Converting PDFs directly to high-quality Markdown for LLM consumption.
Financial Report Analysis: Extracting tabular data from annual reports and statements with high fidelity.

Strengths¶

Vision-Aware: Uses advanced vision models to understand document layout better than traditional OCR.
Markdown Output: Optimized for LLMs, preserving hierarchies and table structures in clean Markdown.
Cost Optimizer: Automatically routes simple pages to cheaper tiers while keeping complex pages on premium tiers.
Ecosystem Integration: Seamlessly connects with LlamaIndex for end-to-end RAG development.

Parsing Tiers¶

LlamaParse offers four tiers that trade off cost, latency, and accuracy:

Tier	Best For	Cost (Credits/Page)
Fast	Plain text, single column, no tables.	0.5
Cost Effective	Text with simple tables; clean markdown.	3
Agentic	Scanned pages, multi-column, charts.	10
Agentic Plus	Dense financial reports, mission-critical accuracy.	45

Limitations¶

Cloud Dependency: Primarily a cloud-based service, which may not suit air-gapped environments.
Latency: High-accuracy vision-based parsing can be slower than simple text extraction.
Cost at Scale: Beyond the free tier, it operates on a credit-based system that can become significant for massive datasets.

When to use it¶

When traditional PDF parsers fail on complex layouts or tables.
When you want "LLM-ready" Markdown output without manual cleaning.
When you are already using the LlamaIndex framework.

When not to use it¶

For simple, text-only PDFs where PyPDF2 or marker would be faster and cheaper.
If your data cannot leave your local environment (though local versions are evolving).

Licensing and cost¶

Open Source: No (Proprietary service)
Cost: Freemium (1,000 pages per month free)
Self-hostable: Limited (Docker image available for enterprise)

Getting started¶

Installation¶

pip install llama-parse

Basic usage¶

import os
from llama_parse import LlamaParse

# Set up the parser
parser = LlamaParse(
    api_key="llx-...",  # can also be set via LLAMA_CLOUD_API_KEY env var
    result_type="markdown"
)

# Parse a document
documents = parser.load_data("./my_document.pdf")

# Access the content
for doc in documents:
    print(doc.text)

Agentic Tier Example (Vision-Aware Parsing)¶

The agentic tier uses advanced reasoning to handle messy layouts and tables.

from llama_parse import LlamaParse

parser = LlamaParse(
    api_key=os.environ["LLAMA_CLOUD_API_KEY"],
    result_type="markdown",
    parsing_instruction="""
    This is a financial report with complex tables.
    Please extract all tables into clear Markdown format,
    ensuring that nested headers are correctly represented.
    """,
    gpt4o_mode=True, # Use GPT-4o vision for maximum accuracy
    premium_mode=True, # Required for Agentic tiers
)

# Using the sync parser for high-priority documents
documents = parser.load_data("complex_report.pdf")
full_markdown = "\n\n".join([doc.text for doc in documents])

with open("output.md", "w") as f:
    f.write(full_markdown)

CLI examples¶

LlamaParse is primarily used via its SDKs or REST API. However, it can be triggered from the LlamaIndex CLI if integrated into a RAG pipeline.

# Example of using a LlamaIndex RAG CLI that might use LlamaParse internally
llamaindex-cli rag --files "./data/*.pdf" --parse-tier agentic

API examples¶

# Create a parse job using the REST API (v2)
curl -X POST 'https://api.cloud.llamaindex.ai/api/v2/parse' \
  -H 'Content-Type: application/json' \
  -H "Authorization: Bearer $LLAMA_CLOUD_API_KEY" \
  --data '{
    "file_id": "cafe1337-e0dd-4762-b5f5-769fef112558",
    "tier": "agentic",
    "version": "latest"
  }'

# Retrieve results (Markdown)
curl 'https://api.cloud.llamaindex.ai/api/v2/parse/{job_id}?expand=markdown' \
  -H "Authorization: Bearer $LLAMA_CLOUD_API_KEY"

# List completed jobs
curl 'https://api.cloud.llamaindex.ai/api/v2/parse?page_size=10&status=COMPLETED' \
  -H "Authorization: Bearer $LLAMA_CLOUD_API_KEY"

Sources / references¶

Contribution Metadata¶

Last reviewed: 2026-05-11
Confidence: high