microsoft/markitdown
Python tool for converting files and office documents to Markdown.
markitdown finally solves the doc-to-LLM problem. Plus 4 more repos that earned a star.
Turn what you learned into a concrete stack decision.
Want the shortlist in your inbox?
Subscribe for the weekly brief that turns new AI noise into the few tools and workflows worth testing.
Python tool for converting files and office documents to Markdown.
Guide
My 5-Minute Crypto DD Stack (markitdown + Claude)
Convert any whitepaper to markdown, feed it to Claude, flag the red flags fast.
Guide
The File Tool Every Crypto Researcher Is Sleeping On
Stop copy-pasting whitepapers into ChatGPT. There's a better way.
Guide
2 Weeks of AI Tool Trends: The Infrastructure Layer Is Winning
Agents are the headline. Infrastructure is where the real momentum is.
Collection of awesome LLM apps with AI Agents and RAG using OpenAI, Anthropic, Gemini and opensource models.
📚 《从零开始构建智能体》——从零开始的智能体原理与实践教程
Beautiful, Modern & Opinionated Linux
Truly independent web browser
Every week I sweep GitHub for repos that actually solve real problems — not just academic demos with 10k stars from a viral tweet. Here's what made the cut this week.
This is the repo from the tweet. And yeah, it deserves the top spot.
markitdown does one thing: it takes any document — PDF, Word, Excel, PowerPoint, HTML, images, audio — and converts it to clean Markdown. One command. No wrestling with libraries. No writing custom parsers for every file type you encounter.
Why does this matter? Because Markdown is what LLMs actually want. If you're building any kind of AI workflow that ingests documents — contracts, reports, slides, invoices — you've been doing the manual conversion dance. markitdown ends that. Feed it a messy PDF, get clean structured text back, pipe it straight into your prompt.
It's from Microsoft Research, which means it'll get maintained. It also handles edge cases most tools miss: tables, headers, embedded images (via OCR). I've been waiting for something this clean for about two years.
"Awesome" lists are usually filler. This one is different.
It's a curated collection of working LLM app examples — not just links, actual runnable code. RAG pipelines, multi-agent setups, voice apps, memory systems. Everything is organized by use case, not by model or framework, which is how you actually search for this stuff when you're building.
What makes it useful: the examples are minimal. No 500-line boilerplate. You can read the code in 10 minutes and understand the pattern. That's rare.
If you're exploring what's possible before committing to a direction — or explaining AI capabilities to a client — this repo is worth bookmarking.
DataWhale is a Chinese AI education community, and this repo is their introduction to building AI agents. Structured learning path, hands-on examples, covers the core concepts (tools, memory, planning, multi-agent coordination) without drowning you in theory.
The caveat: a lot of the content is in Chinese. If you can read it, it's one of the better structured agent learning resources out there — the community produces quality material. If you can't, you're running everything through a translator, which still works but adds friction.
Worth starring if you're trying to understand agents at a foundational level, not just copy-paste from LangChain docs.
This one breaks the AI theme, and I'm including it anyway.
Ladybird is an independent browser — not built on Chromium, not Safari's WebKit, not Firefox's Gecko. It's building its own rendering engine from scratch. That's been considered basically impossible for the last decade.
Why does this belong in an AI tools roundup? Because the browser is increasingly the runtime for AI-generated interfaces. Everything is moving toward AI agents that operate in browsers. A world with only one real browser engine (Chromium) is a fragile dependency for that future.
Ladybird is early — it's not ready for daily use. But the fact that it exists and has serious momentum is worth knowing about. Star it now, check back in 18 months.
DHH's opinionated Arch Linux setup. It's a fully configured development environment — i3 window manager, specific apps, specific keybindings, the whole thing — packaged as a one-command install.
This is not for most people. If you're on Mac or Windows, keep scrolling.
But if you've ever spent a weekend configuring a Linux box and thought "someone should just package the right decisions" — this is that. Basecamp runs lean and ships fast. Their dev environment reflects that. Worth studying even if you don't adopt it wholesale.
markitdown is the only one I'd call urgent this week. If you're building anything that touches documents and AI, install it today. The others are worth a star and a closer look on your own timeline.
Next week I'll be looking at anything that ships around the big inference benchmark drops — expect that list to skew more technical.
Written by McKlaud AI. Want to know which AI tools actually fit your business? Get a free AI audit.