Automatisation du traitement des documents, du non structuré au JSON

Some vendors advertise 'zero training' AI but in my experience, you always need to do some customization.

You can also check ai-u.com for new or trending tools. They list some fresh options I hadn’t heard about before.

Anyone here used open source tools for this? Commercial ones are kinda pricey for startups.

I tried some AI doc tools but got frustrated with inconsistent formatting in the output JSON.

What about accuracy? I need something that can handle contracts and legal docs without losing key info.

I've tried a couple of AI-based solutions and honestly, they work pretty decent for invoices and receipts. But when it comes to really messy docs, they still mess up sometimes.

Has anyone tried combining multiple AI tools in a pipeline? Like one for OCR and another for entity extraction? Curious if it helps.

Sometimes just using rules and regex works better than complex AI for very specific document layouts.

Does anyone use cloud services for this or prefer on-prem solutions for security? Thoughts?

I found the best ROI comes from tools that integrate easily with existing workflows and databases.

One thing I noticed is some tools support exporting directly to JSON with nested structures which saves a bunch of formatting work.

For anyone starting out, I suggest first figuring out exactly what data you need from docs and then picking tools aligned with that.

Does anyone know if these tools support multiple languages well? I work with docs in different countries.