Reverse-engineer a book file format to search it with Claude
After the RAG indexer crashed on an uploaded book, Claude used Unix commands to work out the file format and grep the book manually.
Done with Claude Code
What they did
The commenter uploaded a book into the project context on Claude's web chat. The RAG indexer crashed on the file. Claude then used a series of Unix commands to work out the book's file format, and searched the contents with grep by hand. The source doesn't say whether the commenter prompted any of these steps.
How it went
Claude got around the crashed indexer and could search the book anyway. The commenter found this surprising. The source gives no details on accuracy, the file format, or how long it took.
Worth knowing
Listed under Research & Knowledge. No detailed results are published for this one - verify against the source before relying on it.
Try it yourself with Claude Code
I have a file at [path to file] that my tools can't read properly, and it is a [book/document, format if known]. Use command-line tools to inspect its structure, work out the format, and extract the readable text into a plain text file. Then search it for [topic or phrase] and show me the matching passages with their locations. Don't modify the original file.
Source on HN - Aug 31, 2026
Read the original