Skip to main content Toggle navigation menu Toggle navigation menuArticlesPodcastsVideosTweetsSourcesNewsletters

⌘K

Change language Switch ThemeSign In

Narrow Mode

New Claude Code Skill for Deep Document Research with Word-Level Citations

New Claude Code Skill for Deep Document Research with Word-Level Citations

 ### Jerry Liu

@jerryjliu0

I built a Claude Code skill that allows it to generate a deep research report over any collection of complex docs (PDFs, Word, Pptx)….and generate word-level citations and bounding boxes directly back to the source! 📝

Check out “/research-docs”.

• It parses out text and bounding boxes from every doc with liteparse, in seconds.

• It then generates a full HTML report of the outputs that let you see word-level citations in each page.

Raw Claude obviously has deep research capabilities, but it lacks an audit trail back to the source. This skill gives you a researched report that can be audited by others.

Check it out:github.com/jerryjliu/lite…q

LiteParse:github.com/run-llama/lite…8 Show More

!Image 3: 视频缩略图

01:13

Apr 7, 2026, 5:10 PM View on X

8 Replies

8 Retweets

83 Likes

4,671 Views  Jerry Liu @jerryjliu0

One Sentence Summary

Jerry Liu released a Claude Code skill that uses LiteParse to generate research reports from complex documents with precise word-level citations and bounding boxes.

Summary

Jerry Liu, CEO of LlamaIndex, introduced a new skill for Claude Code called '/research-docs'. This tool addresses the lack of audit trails in raw LLM research by parsing complex documents (PDFs, Word, PPTX) using LiteParse. It generates a full HTML research report that includes word-level citations and visual bounding boxes mapped directly back to the source material, ensuring high verifiability for AI-generated insights.

AI Score

87

Influence Score 42

Published At Today

Language

English

Tags

Claude Code

LiteParse

RAG

AI Agents

Document Parsing HomeArticlesPodcastsVideosTweets

New Claude Code Skill for Deep Document Research with Wor...