Willis: A Docling + RAG Workflow Researching Documents (Updated) #2347
Replies: 3 comments 4 replies
-
|
this is cool @dkylewillis - do you intend for it to be open-source/public? The link in your post is broken – you may have the repository settings on private (not sure if that is intentional). |
Beta Was this translation helpful? Give feedback.
-
|
hi @dkylewillis, this project is rad! I have been looking for a similar problem to index pdfs of manuals and safety regulations in my domain. I think your project suits me here. I will take a look into your repo and maybe in the future will add the contribution :) |
Beta Was this translation helpful? Give feedback.
-
|
I've spent the last two months building a full-featured product, and I'm finally at the point where I’m testing it live. I'm looking for a few beta testers who’d be interested in trying it out and providing feedback. You can check it out at https://www.willis.us, and I’ve included a short 40-second demo video showing it in action. If you’re interested, please email me at dkyle.willis@gmail.com. |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
In my profession, due diligence is arguably the most critical step in meeting deadlines and budgets. This phase requires digging through ambiguous, inconsistent, or poorly specified ordinances to find the one sentence that—if missed—could cost tens of thousands of dollars in redesign. In addition, it’s not uncommon to have more than 5,000 pages of technical manuals defining the specifications that apply to a single project.
Based on the current capabilities of AI, this felt like a natural workflow to improve. Context-engineering techniques such as RAG can provide extremely powerful semantic search when used properly. Package that into a tool and connect it to modern frontier AI models, and you end up with a very capable document-intelligence assistant.
This is exactly how Willis is built. It uses a RAG workflow with Docling to convert documents into structured text, then embeds, chunks, vectorizes, and stores that content. Docling also made it easy to integrate features like visual grounding, which is critical for me if I’m going to use a tool in a professional setting.
Feel free to try it out for free. Also, message me if you’re interested in helping test it—I’m happy to remove usage limits in exchange for constructive feedback.
demo.2.mp4
Beta Was this translation helpful? Give feedback.
All reactions