Glimmer

TLDR:

If you've ever dealt with massive PDFs (i.e. over 1,000 pages), you know how broken search can be. We know people in construction, law, finance, and healthcare who spend hours doing this every day. With Glimmer, you can search PDFs in natural language and get answers with verifiable page sources. Start uploading your own PDFs for free at withglimmer.com.

Hi! We’re Praneeth and Arman and we’re building Glimmer, a new way to search massive PDFs using AI.

The Problem:

Searching through huge PDFs sucks!

Meet James, a project manager at a commercial HVAC subcontractor that does construction projects for the City of San Francisco. The worst part of his job is spending hours every day searching through an 8000-page construction specification trying to figure out things like “What is the fire damper manufacturer?” or “What is the motor speed for HVAC fans?”. Currently, the entire industry uses “Cmd-F” keyword search and manually goes through thousands of search results to find answers to their questions.

When we first met James at his office, he was trying to feed his document into ChatGPT but he couldn’t even upload it – his PDF was over 200MB!

James’ experience is far from unique. In law, lawyers and paralegals read and write long legal papers and contracts, searching for specific details that help them advise clients and do legal work. In healthcare insurance and coding, employees read through massive manuals and patient records to get the right information for billing and handling insurance claims.

✨ The Solution

Glimmer uses AI to make finding information in large documents 10x easier. When a user loads in their PDF (up to 10k pages), we intelligently chunk up their documents so they’re able to use natural language to ask questions and find exact sources within their documents.

Try the interactive demo.

Get Started

Glimmer is completely self-serve, free to use, and works on any kind of PDF. Visit withglimmer.com and upload your first PDF. You can immediately begin searching for anything in your document in natural language.

About us

We both met during our freshman year at UIUC and built several viral generative AI apps together that have been used by 50k+ people. Arman has previously worked on Mercury’s corporate credit/debit cards, helped launch the first-ever BNPL Debit card at Affirm, and worked on AWS’s Heroku competitor called App Runner. Praneeth built the first-ever AI video editing tools for Facebook Reels at Meta and developed image annotation tools for computer vision workflows at Palantir. Now, we’re building the best search engine for PDFs.

This is us at a $3B+ construction site where James’ company works in SF.

Our Asks

• Upload your own PDFs and start searching with Glimmer
• We’re looking for introductions to folks who deal with large PDFs on a daily basis across construction, finance, law, etc. Feel free to reach out to us at founders@withglimmer.com