I download a lot of academic papers and reports, and my downloads folder was a mess of files like `1234.5678.pdf` or `whitepaper_v2_final.pdf`. Manually renaming them was a huge time sink.
To solve this, I built Ramener. It's a native macOS app that: 1. Reads the first few pages of a PDF. 2. Uses an LLM (Aliyun's qwen3-omni-flash) to extract the title, source, and date. 3. Renames the file to a clean `YYYY-MM-DD_Source_Title.pdf` format.
The main goal was to make it feel like part of the OS. It's a standalone app, not a web service. You can integrate it directly into Finder's toolbar or Quick Actions menu, so you can rename files with a single click without leaving Finder. It also has a CLI for power users and scripting.
On first launch, a native settings window pops up for your API key, which is then stored locally.
It's built with Python and packaged with PyInstaller. The code is open source, and I'd love to get your feedback on the workflow, the integrations, or any bugs you find.
GitHub link: https://github.com/jollychang/ramener
k310•3mo ago
Almost always, I can just rename one with its Title, and that's usually the first line.
A summary (usually the abstract) could accompany the file, too long for a tag.
Think email. I'll often send a pdf with its title as the subject and the abstract as the message body above the attachment. Richer than the finder (and let's not gripe on that monster)
I'm also interested in a non-AI, application/image descriptor for my countless anonymous image files, though even a tag would be better than staring at endless screens of titles, or even thumbnail icons.
Again, no slight. I am just interested in something not reliant on AI.