Tool

  • PDF & image to Markdown
  • What is chunks.md?
  • About

OCR models

  • Umi-OCR online
  • Manga OCR online
  • DeepSeek OCR online
  • GLM-OCR online
  • PaddleOCR-VL 1.6 online
  • Granite-Docling online

Guides

  • Extract text from a screenshot
  • Combine documents for AI
  • Add knowledge to your AI
  • Fix flagged Suno uploads

OCR runs on your device — files never leave your browser. Made with ♥ from Yokohama.

← Back

Combine your documents into one AI file

AI knowledge bases cap how many files you can add. Turn a pile of PDFs, Word docs, and notes into one clean markdown file your AI can actually use — converted and merged on your device, nothing uploaded.

Why one file beats a folder

Every AI knowledge base has a ceiling. ChatGPT Projects allow 5 files per project on Free and 25 on Plus. Gemini Gems cap out around 10. Custom GPTs and NotebookLM have their own limits. Hit the ceiling and you're stuck deciding which documents to leave out.

There's a better move: merge everything into one clean file. A single markdown document with a clear heading per source stays under any file limit, keeps the original document names for citing, and gives the model tidier text to retrieve from than a stack of raw PDFs with broken layouts.

Convert anything to markdown first

An AI reads clean text far better than it reads a PDF's layout or a Word file's XML. So before you merge, each source becomes markdown:

  • Word, HTML, text, and CSV — the consolidator converts .docx, .html, .txt, and .csv to markdown right in your browser.
  • Scanned PDFs and images — run them through OCR first to pull out real text, then bring the markdown across.
  • Screenshots — extract the text from a screenshot and add it to the pile.

The result is uniform markdown, no matter what the source was — the format every AI assistant handles best, with several OCR models behind the scans.

Merge into one knowledge file

Once your sources are markdown, the consolidator stitches them into a single document. Add a heading for each file so the AI (and you) can tell sources apart, drop a separator between them, reorder as needed, and export one .md file.

That file is your knowledge base. Upload it to ChatGPT, Claude, Gemini, or Perplexity as a single source — under the file limit, cleanly structured, and ready to retrieve from.

It all stays on your device

Consolidating documents usually means handing a folder of private files to some website. Not here. The conversion and merging run locally in your browser — the same on-device approach as the OCR. Contracts, research, notes, client documents: they never leave your machine, and there's no account or upload step.

If a "combine files for AI" tool asks you to upload your folder, those documents are now on someone else's server. Local consolidation is the version where that never happens.

How to build one AI knowledge file

  1. Open the MD Chunks Consolidator tab.
  2. Drop in your Word docs, HTML, text, CSV, or markdown files — they convert to markdown automatically.
  3. For scans and screenshots, OCR them in the Local OCR tab first, then add the markdown.
  4. Turn on a heading per file, reorder the list, and press Merge.
  5. Download the single .md file and upload it to your AI knowledge base.

Common questions

How do I combine multiple files into one for ChatGPT? Convert each file to markdown, then merge them into a single file. chunks.md does both in your browser: OCR scans and images, convert Word and HTML to markdown, then consolidate everything into one .md file you upload as knowledge.

Why merge documents instead of uploading them separately? AI knowledge bases limit how many files you can add — ChatGPT Projects allow 5 files on Free and 25 on Plus, Gemini Gems around 10. One well-structured markdown file with a heading per source stays under the limit and gives cleaner retrieval than a dozen raw PDFs.

Can I combine Word documents and PDFs together? Yes. Convert Word (.docx), HTML, text, and CSV to markdown in the consolidator, OCR scanned PDFs and images in the OCR tab, then merge the markdown from all of them into a single file.

Is there a file limit on ChatGPT Projects or custom GPTs? Yes. ChatGPT Projects allow 5 files per project on Free and 25 on Plus (40 on Pro, Business, and Enterprise), Gemini Gems around 10, and custom GPT knowledge around 20. Consolidating into one file is the standard way around it.

Does my data stay private? Yes. Every step — OCR, Word-to-markdown conversion, and merging — runs locally in your browser. Your documents are never uploaded to a server.

Related reading

Add knowledge to your AI

Turn any PDF or book into AI knowledge. Extract markdown with chunks.md, then load it into ChatGPT Projects, Claude Projects, Gemini Gems, or Perplexity Spaces.

Extract text from a screenshot

Pull editable text out of any screenshot in your browser. Free, no signup, and the image never leaves your device — OCR runs fully on-device, output is clean text or markdown.

What is chunks.md?

PaddleOCR v5, SmolDocling, MangaOCR, DeepSeek-based Unlimited-OCR — 8 OCR models running in your browser. PDFs and scans to markdown, zero uploads.

Open the consolidator →