LlamaParse Agent Skill

A coding agent skill for LlamaParse can be installed using Vercel’s skills utility. This gives your coding agent the advanced parsing capabilities of LlamaParse, enabling it to process complex documents with different formats, or that contain images, charts, and tables, with a deeper level of document understanding than raw text extraction alone.

Installation

Add the LlamaParse skill to your project:

npx skills add run-llama/llamaparse-agent-skills --skill llamaparse

This downloads a skill file that compatible coding agents (Claude Code, Cursor, etc.) will automatically pick up.

Once configured, your agent will be able to write and execute TypeScript scripts using @llamaindex/llama-cloud to process documents. This means you can have your agent parse and understand unstructured documents on the fly as part of its reasoning process.

Requirements

You will need Node 18+ in your environment, along with a package manager (such as npm or bun) capable of installing the @llamaindex/llama-cloud library and executing TypeScript scripts. You will also need a LLAMA_CLOUD_API_KEY set in your environment.

Example prompts

Once the skill is installed, you can ask your coding agent things like:

• “Parse this PDF and extract the text as markdown”
• “Extract text from all the DOCX files in the ./contracts folder using LlamaParse’s cost-effective mode”
• “Parse all the presentations in the current folder, extract their markdown text, and save any images to the ./presentation-assets folder”