Slashy: AI agents for the inbox
Harsha Gaddipati, cofounder and CEO of Slashy, sprained both of his wrists in February. He was playing basketball in loafers, go figure. He couldn't type for over a month; luckily, the email client he had built could cover for him. "I was in immense pain, but I didn't miss a day of work,"he said.
Slashy is an AI-native email client based in San Francisco and backed by Y Combinator, where it went through the Summer 2025 batch. Email is one of the most obvious use cases for AI, which is why so many have taken a run at it. Gmail got Gemini a while back; Outlook got CoPilot; Grammarly bought Superhuman.
However, Slashy goes far beyond the grammar fixes, quick summaries, and one-sentence reply suggestions of the existing clients. Slashy books meetings for you, drafts replies in full based on your writing style, tracks who still owes you a follow up, and lets you handle things from iMessage or Slack without opening your inbox at all. It does most of it without being asked.
But Harsha's ambition for Slashy is bigger than the inbox. He thinks this is what building AGI looks like at the beginning.
“Email is the starting point,” he said. “We want to make a system that’s so powerful that it can reduce the complexities of work in a way that makes life more enjoyable and work more productive.”
One club, three students, five companies, one goal
It took some twists and turns to get here. Each of Slashy’s cofounders built startups on their own, left them behind, started one project together, pivoted, pivoted again, got into YC, pivoted a third time, and have spent the past year building Slashy.
Harsha started coding because of cricket. During COVID, he went looking for a general manager game for the sport, like what already exists for football or basketball, where you run the roster instead of playing the matches. There wasn't one, so he pulled up tutorials online, taught himself to code, and made it.
"I think that's the nicest thing about being able to code; if something isn't there, if you have this belief in yourself, you're able to create it,” he said. “The only restriction left in life is what you can imagine."
By the time the three cofounders met for the first time in the Startup Exchange Club at Georgia Tech, they each had startups of their own. Harsha was working on a dating app called Bubble Date. Dhruv Roongta, now CTO, had an IT startup and organized hackathons back home in Bangalore. Pranjali Awasthi, the youngest, founded Delv AI at 16, a service to help researchers extract and summarize data from academic papers and online documents, and had funding and press for it by her freshman year. She left Slashy in June 2026.
"All of us wanted to find new cofounders and do a different thing," Dhruv said. "It was a kind of a match made in heaven."
After joining forces, their first project was called Dash. It went through a couple iterations, but the first version of Dash was an AI website builder.
"We were doing well. We got up to 10,000 users," Harsha said. "But everyone churned. The thing with an AI website builder is that you can make a website for your resume, you can make a website for a friend's birthday, but these are all one-off use cases. People just didn't have enough reason to come back."
So they pivoted. Dash evolved into an orchestrating agent for work, much like OpenClaw. It could take action across spreadsheets, calendar, Notion, Slack. That’s when they decided to apply for YC.
They heard back during the summer. Harsha was home in Wisconsin, Pranjali was in Florida, and Dhruv was in Germany. “But we had told YC that we were already in San Francisco,” Harsha said. “So they emailed us like, ‘Can you come tomorrow for an in-person interview?’ We all dropped everything and booked flights.”
Harsha left so quickly, he didn’t even say goodbye to his mom before she got home from work.
By that point, they had raised $30k on their own. They ended up spending over $5k of it just on the flights to San Francisco. When the three arrived the next day, sleep deprived, they dropped off their things in a one-star motel in Chinatown. “It was $70 bucks a night. There wasn’t running water in our room, and we shared a shower with everyone in the hotel.”
But they made it, and clearly it went well. “During our interview, one of the interviewers asked where we were staying. Once we explained we were in a one-star hotel, they said ‘hopefully we can change that for you very soon.’” They were accepted in the YC Summer 2025 batch; even though Dhruv and Pranjali still had classes to finish at Georgia Tech, they put school on hold and moved to California.
Everything starts with an email
After being accepted into YC, the trio decided to refocus their efforts on people. They wanted to create a company that would save time and minimize stress. They quickly landed on email.
“Email is one of the most common functions humans do at this point, which is a little bit weird when you think about it, right?” Harsha said. “It’s probably in the top five to ten actions, after sleeping and eating.”
"We realized that AI will be the future of work, but it's still going to be part of an interface," Harsha continued. "We looked at email and calendar, and we were really frustrated with the existing solutions – Gmail, Superhuman. We were like, 'This should not be how the state of software looks in 2025.'"
They pivoted into email after YC Demo Day and renamed the company Slashy that August.
Slashy today is an email UI that connects an AI agent to everything in your inbox and calendar.
- It learns from your messages, so it can start drafting follow-ups for you.
- It connects with your calendar app and Slack to schedule meetings, then texts you beforehand with a brief on what the meeting is about.
- Custom labels get assigned automatically by message type: sales outbound, team members, event invites.
- Over time it learns what you think is important and what can go to the back burner.
"The bar for an email client is high," Dhruv explained. "You need a full-featured web app, you need a mobile app, a desktop app. All of it needs to sync perfectly, all of it needs to look good, you need to have the keyboard shortcuts. It takes a lot of building effort initially."
They looked to some of the disruptors in UI, such as Notion and Linear, and sought to create the email equivalent. "They all succeeded not just because the product is good, but because it's very nice to look at and use. The product needs to be delightful," Dhruv said.
Making it pleasant to look at is one problem. Making it think is another, which is Dhruv's domain. He got into coding around eighth grade, when a C++ class at school bored him and he started teaching himself Python and machine learning. In college he interned every semester, ranging from more machine learning, to inference, to data science.
"I got exposure to front-end stuff, back-end stuff, AI/ML stuff, stuff where you're pushing directly to prod, stuff that's more research," Dhruv said. "I got to see all the aspects. I don't go super deep on one thing, but I will say, I started building agents very early on, like summer 2023."
The agents Dhruv had been building became the machinery inside Slashy. The automated email labeling system was one of the first pieces they built, sorting mail into buckets like important, customer, FYI, and billing, and it learns from every correction the user makes. Dhruv says it’s still his favorite feature.
"Without that, the inbox is just so full of emails that it's literally impossible to figure out what I need to get done," he said. "With the labeling, it's so much easier because I know exactly which emails I need to respond to."
The future of work
Everyone in AI wants to be the one to build AGI. The pitch is the next trillion dollar company or the dawn of a new economic age where our previous assumptions about life, work, and productivity fall away.
Harsha thinks people are thinking about AGI all wrong.
"Me and my co-founders' hot take is that we think AI is smarter than almost anyone already," Harsha said. "Everyone right now thinks that the problem is that the models aren't smart enough. I think it's the interfaces for AI that are the problem, not the models."
To the Slashy team, AGI is real when two things become true: first, the models stop making the little mistakes, and second, the models become truly capable of handling the tedious parts of life on their own without a person in the loop.
The best testing ground for that is email. It's where people already coordinate their lives, interact with each other through text, make big decisions, and lose an enormous amount of time doing it.
"People need to think from a more people-first approach. What makes the smartest human smart?" Harsha continued. "We try to model the agent like a smart human."
If the ultimate goal is AGI, Slashy is betting that it shows up in the inbox before anywhere else. It’s not about creating the smartest model, but building the model that acts just as well as you do in the parts of your day you'd rather skip.
Slashy is one of many startups building with Slash. We're proud to support founders from day one with the financial tools they need to grow. If you're building something, we'd love to work with you. Get started today at slash.com/startups.
Apply in less than 10 minutes today
Join the 10,000+ businesses already using Slash.
