Clerve

This is the Document Vault which stores the forms for different Clients based on their name for attorneys to access later.
This is the dashboard a user will see when in the process of creating a new case or opening an existing one.

Inspiration

Immigration law is one of the most paperwork-heavy and time-consuming legal domains. Working directly with lawyers, we observed that repetitious data input can take five to seven hours every case, but legal strategy requires less than ninety minutes. We created Clerve in response to these inefficiencies, allowing attorneys to concentrate on their clients rather than paperwork.

What it does

Clerve automates the process of converting structured client data into completed USCIS immigration forms. By reducing hours of manual data entry into a streamlined 10–30 minute workflow, it significantly boosts efficiency and reduces operational costs for law firms.

How we built it

We combined full-stack development with machine learning to create Clerve. Data is extracted from PDFs, transformed into structured language, and then immediately mapped into USCIS forms using our pipeline. A simple user-friendly user interface created especially for lawyers supports the system, guaranteeing accuracy and usability.

Challenges we ran into

One of the biggest challenges was handling inconsistent and messy input data from PDFs and legal documents. Ensuring high accuracy in extraction while maintaining compliance with strict legal requirements required multiple iterations of our ML models and its validation logic.

Accomplishments that we're proud of

We built a product that directly addresses a real-world pain point validated by legal professionals, and a solution that can save thousands of dollars per lawyer annually.

What we learned

We learned that in regulated industries like law, accuracy and trust matter more than speed alone.

What's next for Clerve

Next, we plan to expand Clerve into a full automation platform with multi-form support, AI-powered compliance checks, and a unified client data layer.

Built With

Submitted to

Los Altos Hacks X

Created by

I worked on the form generation for each USCIS form. This involved creating specific form fields for each document to make it fillable. Then a mapping engine needed to be created so that the code will scan the questionnaire and scan the document while matching the field keys to paste the user info onto the document. Then it will automatically download. I also worked on the document organization architecture so that the forms that are generated are saved in the client's specific folder for the attorney to access later.

Saurya Bachala
I made the entirety of the UI, made the risk assessment, the form generation didn't work so I have spent like the last 5 hours working on that, I integrated the extraction with the API with the questionnaire, made sure that the questionnaire asked all of the quesitons, tried to get the form to populate correctly, and made the whole pitch deck. I also came up with the idea

Adi Anand
Elamaran Elango