How to Set Up a Claude Projects Knowledge Base That Actually Works
A step-by-step guide to preparing documents for Claude Projects so answers stay accurate, grounded, and within the project knowledge limit.
Why structure matters
Claude reads project knowledge as context on every message. Messy files mean every question pays for the mess: repeated boilerplate, broken tables, and navigation menus all compete with the content you actually care about.
The preparation workflow
Start by converting everything to text in one pass with the Batch Document Extractor. It handles PDFs, Word files and images together and returns one clean file per input.
For web pages and HTML exports, the Markdown Converter strips scripts and layout while preserving headings and links.
- Name files by topic: pricing-policy.md beats doc_final_v3.md.
- Put a one-line summary at the top of each file.
- Remove duplicate versions of the same document.
- Check the total in the Token Counter before uploading.
Write good project instructions
Tell Claude what the files are and how to use them: which file is authoritative, what to do when files disagree, and whether to cite filenames. Two or three sentences of instructions often improve answers more than adding more documents.
Try it free in your browser
- Markdown Converter — Paste HTML, DOCX, or webpage text and get tidy Markdown.
- Token Counter — Live token, character, and context-window meter for any prompt. Uses official tiktoken encoders.
- Batch Document Extractor — Drop PDFs, images, Word docs, and text files together. Get clean text out — combined or zipped one-per-file. Browser-side OCR included.
Tools mentioned
Frequently asked
What file format is best for Claude Projects?
Plain Markdown. It keeps headings and lists, which Claude uses to understand structure, without the noise of PDF layout or HTML tags.
Should I upload one big file or many small ones?
Many focused files with descriptive names. Claude can cite and navigate them more reliably.
Keep reading
How to Prepare PDFs for ChatGPT, Claude, and Gemini
A practical guide to extracting clean, AI-ready text from PDFs — born-digital and scanned — so ChatGPT, Claude, and Gemini answer accurately and don't waste tokens on headers, footers, and page numbers.
The Best Way to Feed Long Documents to Claude (and Other Long-Context Models)
Claude's 200K-token context is generous, but you'll still want to clean, compress, and structure long documents before sending them. Here's a step-by-step playbook.
Cleaning Whisper Transcripts for AI Summaries
OpenAI Whisper, Otter, and YouTube transcripts are full of timestamps, filler words, and speaker noise. Here's how to strip them before sending to ChatGPT or Claude — and why it matters.