extend

Last quarter, we closed 6-figures in ARR within 2 weeks from a single tweet. Here's what happened: A few months ago, I tweeted about processing a 91-page PDF with nested tables and handwriting. Something clicked w/ doc processing twitter — 500K impressions and hundreds of signups later, our calendars were overflowing. The phone booth basically became my second home. Every person on the team stepped up to handle demos. We were running on a combo of coffee and pure adrenaline. Here's the thing about sudden growth spurts: they stress test everything. Our sales and onboarding process got pushed to the breaking point, but that pressure forced us to evolve. We now have a completely dialed-in process that can handle 10x this volume. Looking ahead, we have a simple formula to follow: 500K impressions → 6-figures in ARR within weeks....what happens when we scale that up to 100M impressions? You can do the math 📈

Looking through old photos while preparing for our offsite, I found this picture of our first real office. It made me think back to the early days, and how we shipped our way out of 2 pivots... There's only one piece of advice for getting out of pivot hell: focus on validating problems quickly. We moved lightning fast — going from 0 lines of code to onboarding our first customer in 48 hours, then closing a 6-figure ARR deal two weeks later. Looking back, my favorite memories from those early days: - Showing up at a potential customer's office (who was ghosting us) with the best croissants in SF to convince them to use our MVP 🥐 - The day we moved into this office and putting up the "make something people want" poster - Scratching down our initial pricing structure on the back of a napkin over lunch Feeling immensely grateful to the early customers and team that took a chance on us, and couldn't be more excited about the road ahead!

Just wrapped up extend's 4-day offsite in Puerto Rico! After a huge year (ARR & usage volume 📈📈), we got the team together to plan our next chapter of growth. A few highlights: 1/ Ambitious 2025 goals — after 15x'ing revenue last year, we're doubling down on what's working by leaning into technical & AI teams as our core ICP. We also identified a few ways we can pour gas on the fire (stay tuned for a big product announcement later this quarter 🤫). 2/ Reaffirming our culture — I spent time sharing Extend's founding story with the team, reliving the early days from navigating the idea maze to signing our first customer. We talked about the ups and downs of the journey, and how at the end of the day, customer obsession is the only north star we can rely on. 3/ Of course, we couldn't leave without celebrating — we hiked in the rainforest, got morning coffee by the beach, stayed up late playing Catan, and explored Old San Juan. Offsites are a good reminder that, even though we may be heads down shipping features and processing millions of documents, startups at the end of the day are just about people — and this team is the best one I've ever been a part of. Back in NYC and I'm feeling energized and ready for our biggest year yet. Let's get it 🔥 P.S. We're growing quickly and hiring senior engineers in NYC who are excited about building the future of unstructured data processing. DM me if interested!"

At the end of the day, customer feedback like this is what makes the tough moments of building a startup all worth it. Checkr is a legendary YC company, and we're thrilled to have them onboard as a Extend customer. Checkr processes millions of background checks, and they needed a reliable way to extract data from complex documents without constant maintenance overhead. Using Extend, they're able to focus on innovating their core product instead of managing document infrastructure. We're helping more companies like Checkr transform their document operations. If you're dealing with complex documents at scale, DM me and let's chat!

we flew out a video crew to one of our customer's offices here's how Vendr uses Extend to process millions of documents 🔥

We grew revenue 15x last year to 7-figures in ARR with a small team, and we're looking for extraordinary engineers to help us scale even faster. We're hiring for multiple roles in NYC with a $10k referral bonus for successful hires! Here's why you should join us: - We're building the next generation document processing platform — the market for document processing has expanded 1000x due to LLMs, and existing solutions are all low NPS - We have a product loved by users and being used in mission-critical flows at companies like Chime, Flatiron Health, Brex, Opendoor, Checkr, and many more - We're supporting customer and revenue metrics with half the team size of other startups, so everyone joining at this stage will have outsized impact - You'll be joining a talent-dense founding team (think: former founders, world record holders) tackling novel problems at the intersection of LLMs & complex data processing - We have a low valuation with extremely strong traction and metrics, so now is the perfect time to join Link to apply is in the comments below (or DM me directly), and we have a $10k referral bonus for successful hires!

On sales calls, the first question most customers have is "sure, but can it handle *my* difficult docs?". So to that end, I'm starting up a new series of extend vs. difficult PDFs! First up — zero-shot table extraction on a (internal) complex healthcare doc. This example in particular is interesting because it has (1) a table split across pages, (2) content in-between, and (3) no delimiter for rows or columns. What enables this is vision models specialized for tabular structures + semantic chunking that identifies the table continues over multiple pages (and should be treated as one continuous element). If you have complex documents, comment them below and we'll take a crack at it!