The fact that its able to correlate between points of unstructured data well, i.e. in an invoice locate values for fields correctly irrespective of it being written to the right, below or in tabular format. Review collected by and hosted on G2.com.

If there are multiple values at different places in the document corresponding to the same key/field, it should select more sensible or more frequent value which is not the case. Also, while parsing invoice, if there if an email written in the document, but label 'Email' is not there, it won't make sense and parse it as a field correctly. Review collected by and hosted on G2.com.

So I ended up using just the extraction part of LLM whisperer and I used the API to process over 10,000 individual PDFs. The features it has built into the API such as changing the quality type, bounding around tables was super helpful. Review collected by and hosted on G2.com.

The hardest part in my experience was getting the API to connect. The URLs, header names, waht to send and not send in the body were a bit clunky to figure out. Review collected by and hosted on G2.com.

It just works, no need to fine-tune models or wrestle with layout parsers. I can send messy PDFs (including scanned ones) and get clean, structured text back through a straightforward API, which saves me hours of preprocessing. Review collected by and hosted on G2.com.

While the API is fast, detailed documentation on edge cases would help even more. Review collected by and hosted on G2.com.

The task that I was trying to do was made simple by using this software. I was trying to convert pages of hand written data into a format where I could easily transfer to excel. This program made it into an easy and very accurate table that I could use to upload data into excel. I was saved probably hundreds of hours of manual input because of Whisperer Review collected by and hosted on G2.com.

I had to use AI to convert to a table that could be uploaded to excel. Getting the values that were handwritten into text made up for this inconvenience Review collected by and hosted on G2.com.

What I appreciate most about LLM Whisperer is its ability to extract Hindi text from scanned PDFs. This feature stands out as particularly useful for my needs. Review collected by and hosted on G2.com.

After extracting the dataset, I found that the error was minimal. However, since the text was in Hindi, a few data fields were not extracted correctly. Review collected by and hosted on G2.com.

I LOVED how easy it was to use and how precise the extraction was. Implementation and integration was effortless. It had a simple rich set of features that I use frequently. It's so easy to use that I've never had to call customer support. Review collected by and hosted on G2.com.

No issues whatsoever. I love the product. Review collected by and hosted on G2.com.

LLM Whisperer offers a great OCR. Especially, the Layout preservation is a key component in my dissertation project. I tried multiple other systems and did not manage to achieve such great results.

The digital playground allowed me, to test the System with my data before committing to it; a feature I liked a lot, since it showed me the possibilities of the tool and allowed me to properly evaluate the tool in the process of creating my workflow.

It allowed me to create a digitalisation pipeline, to digitise old data (FMRD ancient Coin Data in Numismatics) from table books via RegEX to Dataframes with a great deal of precision.

Since the old tables were only separated by tab stops and had many irregularities, I needed a perfect layout recognition to later use Regex on the OCR Result of my scans.

The fast processing of the scans allowed me to digitise about 40,000 coins in about a month of work.

Also, the ease of use and the possibility to directly integrate it via API are great. Review collected by and hosted on G2.com.

So far, I have not encountered any problems with the software. Sometimes, if the scan is a bit tilted, LLM whisperer struggles a bit with recognition, but that is mainly due to the old books with very thin pages. Review collected by and hosted on G2.com.

Massive PDF conversion by just prompting based Review collected by and hosted on G2.com.

Found some mistakes during PDF document parsing which is hard to find when we process thouands of documents. The JSON output always included metadata for which I reauired in my prompte to not include. Didn't found any solution for that. Also, had some issue to connect the API Review collected by and hosted on G2.com.

We’re very satisfied with the results from LLM Whisperer so far. It delivers well-structured, high-quality input for LLM processing, even when our source PDFs are far from perfect. The integration with Unstract Cloud makes it even better - thanks to line numbering and text highlighting, our human review loop works smoothly and efficiently. Review collected by and hosted on G2.com.

Pricing could be more affordable, but considering the limited number of alternative solutions (which come with their own downsides), it's acceptable. Review collected by and hosted on G2.com.

That ouf of the different softwares for document extraction i got the best use out of this. In the project employed, it serves as a bridge that enables natural-language processing and intelligent responses to user queries about investment funds. Review collected by and hosted on G2.com.

Right now, not much, maybe that the option for the complete extraction is much more expensive Review collected by and hosted on G2.com.