Docling - IBM Library to Make Documents AI Ready - Install and Test Locally

[@concretec0w]

Appreciate the hard work on the channel btw. I don't know how you pump out so much quality content 👍

[@fahdmirza]

I appreciate that!

[@maxjesch]

Concise, hands-on, informative --> great contet

[@fahdmirza]

I appreciate that!

[@peterstaar4912]

Thanks for highlighting our work, feel free to contact us any time! Awesome video!

[@fahdmirza]

Thank you, please also share it among your network. How can I contact you guys ? I was thinking of having a fireside chat about Docling with anyone from your team , if you like, thanks.

[@concretec0w]

This is where the small VRAM , consumer GPUs get the opportunity to significantly boost quality of llm responses more than we've seen in a long time! Very exciting times. I wonder if it can be used to boost agentic function calling capabilities, would open up much improved agentic coding potential in plugins like cline (claude-dev) :)

[@d1m18]

Thank you but in the future I would prefer to have two distinct videos, one that shows install steps, and another that shows what the tool does and how to use it

[@bharatguptag1994bharat]

Can I use it for invoice data extraction like converting the invoice in markdown/json and then feed to LLM ?

[@ROKKor-hs8tg]

This is great, thank you.

[@fahdmirza]

you are very welcome

[@naveennoelj]

Excellent Video and library. You always have something new to show us. Thanks. Qs: How would you vectorize the PDF so that we can build RAG applications. I saw some mention about integration with LlamaIndex and Langchain.

[@fahdmirza]

Thanks for kind words. RAG is certainly a usecase as I mentioned in video. I have various videos covering that part already, please search the channel. Thanks.

[@ajarivas72]

​@@fahdmirza I like your channel very much.

[@NLPprompter]

why why so cute mascot aw.... that cute duck

[@fahdmirza]

indeed

[@tungthanhle6797]

awesome work! Can I run without the GPU?

[@fahdmirza]

Yes, of course

[@AshWickramasinghe]

Can this be used programmatically? Within a workflow of an Agentic app?

[@fahdmirza]

yes definitely, please see towards the end of video, where i show python code. thanks.

[@d.d.z.]

Can you make an example of RAG using this tool?

[@fahdmirza]

I already have few videos around it on the channel, please search.

[@enricollen]

Thanks sir

[@fahdmirza]

So nice of you

[@tecnom7133]

thanks

[@fahdmirza]

You are welcome.

[@shreymishra646]

Docling is slow, for any project that really wants to scale using Docling and waiting for 30-85 seconds per page is extremely slow