Agnos
Download

Agnos vs MarkItDown

Last updated

If you write Python, need text read from scanned pages, work outside Windows, or want conversion built into your own pipeline, MarkItDown is the better pick, and it is free. If you want a folder of text-based PDFs, Office files and web pages cleaned on Windows without code, with a list of what it removed and the rule for each, Agnos is made for that job.

Gidion makes Agnos and wrote this page, so read it with that in mind. Every statement about MarkItDown comes from its GitHub repository (its README, release notes, code and test workflow) and its PyPI entry, listed under Sources.

At a glance

AgnosMarkItDown
What it isA Windows app, made by Gidion, that cleans documents before they go into AI tools and lists what it removed.A Python tool from Microsoft that converts files and office documents to Markdown.
Who it’s forPeople preparing documents for Custom GPTs, NotebookLM, Claude Projects, Copilot, a RAG index or training data, without code.Developers feeding documents to LLMs and text-analysis pipelines. Its README says the project provides Python libraries for other systems, not end-user applications.
Where it runsYour PC. Documents are never uploaded; Agnos goes online only to activate, check or deactivate a licence.Your computer or a server, as a Python package or in Docker. Its README calls built-in document conversion offline, on local compute. The optional Azure converters are cloud calls. The OCR plugin and image descriptions send images to the AI model you connect, and audio transcription uses Google’s online speech recognizer.
Needs codeNo. No Python to install, no terminal. The command line is coming to the Microsoft Store version in an update.Python 3.10 to 3.14 and pip to install. Then a command line, or a Python API for code. Its README leaves desktop apps to separate projects.
File typesText-based PDF, Word (.docx), PowerPoint (.pptx), Excel (.xlsx), web pages, CSV and TSV, JSON and JSONL, plain text, Markdown, XML and other markup or config files read as text, and source code. Old Office formats, images, audio, video and archives are refused.PDF, Word, PowerPoint, Excel, images, audio, HTML, CSV, JSON, XML, ZIP files, YouTube URLs, EPUB “and more”. With the [all] install its README shows, it also reads Outlook messages and older .xls workbooks.
Scanned PDFs (OCR)No OCR. A page with no text layer is skipped and named in the file’s warnings; a PDF with no text at all is refused.With add-ons: the markitdown-ocr plugin reads scanned PDFs page by page through a vision-capable AI model you connect, and Azure Content Understanding offers cloud OCR.
TablesRead from Word, PowerPoint and web pages. In PDFs, only with the slower Layout reading mode. Excel sheets become tab-separated rows, without formulas or formatting.Written as Markdown tables from Word, PowerPoint, Excel, CSV and web pages; its README names tables among the structure it preserves. In PDFs, it writes aligned Markdown tables for pages its code recognizes as holding tables or forms (since version 0.1.5); other pages come out as plain text.
What it outputsA RAG corpus in JSONL (an id, the source and the pages, slides, sheet or headings on each piece), cleaned .txt files, training data with a validation split, or Everything: the dataset with every removed piece and why, a dataset card and the settings used.Markdown, one document per conversion, to the screen or to a .md file.
Shows what it removed, and whyYes. Review lists what it removed and the rule for each, up to 500 removals in a file, and the Everything export includes rejected.jsonl, with every removed piece and its reason.Not described in its README, which is about converting documents rather than removing parts of them.
Price14-day trial of everything in Pro, no card. Then $99 once for one person on up to 2 devices, or $399 once for up to five people at one company, on up to 10 devices, plus tax where it applies. No subscription.Free of charge. The optional Azure converters are billable Azure API calls, and the OCR plugin runs on the AI model you connect, which may be a paid service.
Operating systems64-bit Windows 10 (version 2004 or later) and Windows 11, from the Microsoft Store.Not tied to Windows: it is a Python package for Python 3.10 to 3.14, and the project’s own tests run on Linux and Windows.
LicenceCommercial, under an end user licence agreement.Open source, under the MIT License, copyright Microsoft Corporation.

Where MarkItDown is the better choice

  • Your documents are scans. Agnos has no OCR, so a scanned page gives it nothing to read. MarkItDown’s markitdown-ocr plugin reads text from pictures inside PDF, Word, PowerPoint and Excel files. For a scanned PDF, it sends each page as an image to a vision-capable AI model you connect. Its Azure Content Understanding option offers cloud OCR for scanned PDFs instead. Either way, the pages go to the service you chose.
  • Tables matter. MarkItDown writes tables from Word, PowerPoint, Excel, CSV and web pages as Markdown tables, and since version 0.1.5 it writes aligned Markdown tables for PDF pages it recognizes as holding tables or forms. Agnos reads PDF tables only in its slower Layout mode, and turns Excel sheets into rows of plain text.
  • Complex layouts, or fields to pull out. Its Azure options add cloud layout extraction, and Content Understanding can write fields such as invoice amounts and dates at the top of the output. These are billable Azure calls.
  • Other kinds of file. MarkItDown’s README lists images, audio with speech transcription, ZIP files, YouTube URLs and EPUB books, and with the [all] install its README shows, it reads Outlook messages and older .xls workbooks too. It converts a web page straight from its address, writes PowerPoint chart data as tables and includes speaker notes, reads Jupyter notebooks and RSS feeds, and reads video through Azure Content Understanding. Third-party plugins can add more formats. Agnos reads web pages saved as files, leaves speaker notes out unless you turn them on, does not read charts, and refuses images, audio, video, archives and old Office formats.
  • You don’t use Windows. MarkItDown is a Python package that needs Python 3.10 to 3.14. Its own tests run on Linux as well as Windows, and the project includes a Dockerfile. Agnos runs only on 64-bit Windows 10 and 11.
  • You are building a pipeline. MarkItDown is a Python library. It also has a command line that reads from a pipe, a Dockerfile, and an MCP server, markitdown-mcp, that AI apps such as Claude Desktop can call. Its README says it is meant to be built into other systems, and covers server-side use. Agnos is a desktop app, and its command line is not in the Microsoft Store version yet.
  • You want it free and open. MarkItDown is free of charge under the MIT License, and you can read and change its code. Agnos costs money after its trial.

Where Agnos is the better choice

  • You don’t write code. Agnos is a Windows app from the Microsoft Store. You drop in a folder, check the result in Review and press Build and download, and there is no Python to install: Agnos brings its own. MarkItDown starts with Python and pip, in a virtual environment as its README recommends, and the README leaves desktop apps to separate projects.
  • You need the clutter taken out, not only converted. Agnos removes navigation menus, cookie banners, “enable JavaScript” notices, placeholder text, page numbers, running headers and footers at the edges of PDF pages and slides, and passages repeated within a file or across files. MarkItDown’s README describes converting documents to Markdown; it does not describe removing any of these.
  • You want to see what was removed. Review lists what Agnos removed and the rule for each piece, and Keep this anyway puts a removed piece back. Lines removed while cleaning, such as page numbers and cookie banners, are the exception: they come back only when you switch their setting off and clean the file again. The Everything export adds rejected.jsonl, with every removed piece and its reason. If one file has more than 500 removals, Review counts them all but lists the first 500.
  • Your documents must stay on your PC. Agnos reads, cleans and exports on your computer and has no online options to switch on. It goes online only to activate, check or deactivate your licence, and sends no document text. MarkItDown’s built-in document conversion is offline too, its README says. Its Azure options are cloud calls, its OCR plugin sends images from your documents to the AI model you connect, which may be an online service, and its audio transcription uses Google’s online speech recognizer.
  • You have a folder of mixed files for an AI tool. One project takes text-based PDFs, Word, PowerPoint and Excel files, web pages, CSV and JSON together. Agnos exports a RAG corpus in JSONL, cleaned .txt files for Custom GPTs, NotebookLM, Claude Projects or Copilot, or training data with a validation split. MarkItDown gives one Markdown document per conversion, and its README does not describe splitting the text into pieces or writing JSONL.
  • You’d rather pay once than keep a cleaning script going. MarkItDown is free and Agnos is not. Agnos is $99 once for one person, or $399 once for up to five people at one company, after a 14-day trial with no card. There is no subscription, and every 2.x update is included.

Using them together

MarkItDown writes Markdown files, and Agnos reads .md files, so MarkItDown can go first for files Agnos does not read, such as scans, EPUB books or Outlook messages. Agnos then cleans the Markdown, lists what it removed and builds the export.

  1. Convert with MarkItDown. Save each result as a Markdown file, as its README shows: markitdown report.pdf -o report.md. For scans, add an Azure option to that command, such as --use-cu with your Content Understanding endpoint. Or run the OCR plugin from Python with the AI model you connect, and save result.markdown as a .md file. The plugin’s README shows --llm-client and --llm-model options, but MarkItDown’s command line has neither, and without a model the plugin skips OCR without saying so.

  2. Add the Markdown files to Agnos. Drop the .md files, or their folder, on the Import step of a project with the RAG / Retrieval Corpus purpose, which keeps headings, tables and lists as written.

  3. Review and export. Open Review, choose a file and press Removed, then build the export. In a RAG corpus, each piece carries the Markdown headings it sits under.

Three things to know:

  • For a text-based PDF, give Agnos the PDF itself. Agnos finds running headers and footers at the edges of each PDF page, and records the pages each piece came from. A Markdown file has no pages, so that rule has nothing to look at; Agnos still removes a page marker such as “Page 3 of 10” on a line of its own.
  • If MarkItDown’s OCR plugin or an Azure option sends your pages to an online service, those documents leave your computer at that step. Agnos’s part stays on your PC.
  • MarkItDown’s OCR plugin wraps the text it reads in *[Image OCR] and [End OCR]* lines, and starts each PDF page with a heading such as ## Page 1. Expect those in the Markdown you add to Agnos, and check them in Review.

Both tools are good at their own job. MarkItDown turns many kinds of file into Markdown for code and AI models to use; Agnos cleans a folder of documents on Windows and shows what it took out. If yours is the second job, you can try Agnos free for 14 days, with no card. What it does not do is listed in What Agnos doesn’t do.

Sources

Checked 17 October 2026. MarkItDown has no website or pricing page of its own: its PyPI entry names the GitHub README as its documentation.

MarkItDown

Agnos

MarkItDown is a trademark of its owner. This page is ours, not theirs; tell us at hello@agnos.cloud if anything here is out of date.