After Layoffs Killed His Dream-Job Offer, He Spent Six Weeks Building 'Chat With Your PDF': Launched to 16 Followers, $59K MRR in Three Months, Now About $10M a Year With Zero Funding

Process

To be clear up front: Chatbase now has about 30 employees, so it's long past being a one-person project. It's included because it has only ever had one founder, it has never taken a cent of outside investment, and its start was about as ordinary as it gets: a college student who lost a job offer, and an X account with 16 followers.

Stage 1: The Dream-Company Offer Disappears

Yasser Elsaid is from Egypt and studied computer science in Canada. In 2022, as tech companies began mass layoffs, the new-grad offer he had from his "dream company" was rescinded.

He later wrote on X: "This had a big impact on my mental health and I was about to fall back into depression." But he also made a decision: "I always knew that I want to build a side project to make money online, so I decided that I won't allow my self to be a victim and thought this was the perfect opportunity to start building."

He had been building side projects all along. In school, he did internships while working on his own small projects, such as a course-rating site called Rate My Courses, "with the goal of just having something that can make money." He says he was inspired by indie makers like Pieter Levels.

Stage 2: Spotting the Opportunity Before the ChatGPT API

The ChatGPT API hadn't been released yet; he was playing with OpenAI's earlier Davinci-003 model. "Back then there was no API where you can send the chat and get back a response. It was just you're talking to the base model, just the completion, not chat completion."

He noticed a gap: the model was powerful, but it didn't know your own data. What if you could add a textbook, a book or a website and then chat with it? People later called this approach RAG (retrieval-augmented generation), but back then it didn't even have a name. He saw people writing papers and building command-line tools, "but there was no product that you can open a website, add a document and then chat with it."

He talked to a lot of people about the idea. Many didn't see the value, or saw the value but didn't believe the models would improve. He bet that they would.

Then he made an impulsive decision: "I just stopped everything I'm doing in university. I just stopped going to class." He had a conviction that "this is an opportunity that doesn't come often."

He didn't do market research or user interviews. "In a fast-moving industry like AI, you don't have time to talk to users because there is no users. It's a new paradigm. They've never used something like this."

Stage 3: Launched to 16 Followers, Paid Within 30 Minutes

He was coding by hand, "so it was still slow," and shipped the first version about six weeks later.

On February 4, 2023, he tweeted that he had built Chatbase, which lets you create a chatbot from any PDF document. In the demo, the chatbot had been trained on a document explaining how Chatbase itself worked. The tweet listed his tools, LangChain, Pinecone and OpenAI, deployed on Vercel, and openly credited three other developers as inspiration.

He had 16 followers at the time. The tweet has since passed 800,000 views. The next day, he posted another demo: uploading Naval's book to Chatbase and asking it questions.

He started charging right away. "I had the first payment through Stripe 30 minutes after I put out a pricing page." According to web archives, pricing in March 2023 was $10, $50 and $200 a month.

On May 10, 2023, about three months after launch, he tweeted: "Today, Chatbase is doing $59,000 in MRR." He added: "Losing that job might have been the best thing to ever happen to me."

Stage 4: Customer Calls at 2 a.m.

Growth was too fast for one person. "I had some angry customers calling me at 2 am. That was my whole day, just building and improving the product and talking to customers."

About two and a half to three months after launch, he hired his first person: an engineer friend from school. A month later he hired a second, also a friend. "I didn't have the choice. It was extremely obvious that I need to hire someone. It was just this exponential growth."

He didn't raise money or find a co-founder, partly because there was no time. "I wasn't in SF. I didn't think about what people do when they want to start a company, like first you need a pitch deck, then you talk to investors, then you try to find a co-founder. I just saw the opportunity and I started building."

Looking back, he thinks he hired too slowly. A bootstrapped company spends customer revenue, and "you want to make sure that it's ROI positive all the time," which made him overly cautious early on.

Stage 5: From "Chat With Your PDF" to an AI Customer Service Platform

Chatbase today looks very different from its first version. It has gone from "upload a PDF and chat with it" to a platform for AI customer service agents: businesses give it documents, websites and other data, and it creates AI agents that answer customer questions in a website chat widget, email, voice and more, plus a helpdesk inbox. The entry-level Hobby plan is now $40 a month.

The path wasn't straight. "In AI startups, the inputs change every day. You make a decision based on what you know today, but then you shouldn't have ego. If some inputs changed a week or a month later, you need to go in a different direction very quickly."

He even argues a small company should be a kind of dictatorship: "If you have trust in the founder and you think they're smart enough, then it should be a dictatorship. You don't want them to be slowed down." He also concedes that if the founder isn't smart enough and the team doesn't trust them, the company won't work either way.

He keeps telling the team that changing your mind is good: "Inputs are changing very quickly, and sticking to something just because you don't want to look bad is such a bad idea."

Stage 6: $9M ARR, and Then Spending to Grow

In April 2026, Yasser appeared on the Solo Founders podcast. He said Chatbase had reached $9M ARR in about three years, with about 30 people and no outside investment. According to Stripe data read directly by TrustMRR, as of July 31, 2026 Chatbase had $863,632 in MRR (about $10.4M annualized), $19,213,227 in all-time revenue and 12,424 active subscriptions.

The growth playbook has changed too. Early growth was organic; later the company started investing in paid ads, content and outbound sales, building a sales team only at around $7–8M ARR. "We have been profitable for a long time and now we have a lot of conviction on what we're building," he said. "We may be even spending more money than a lot of VC-backed companies."

He grew up swimming and doing archery, and says individual sports taught him one thing: "you can only blame yourself." It's the same reason he likes being a solo founder: "I like the fact that everything is my fault."

As for why he hasn't raised, he's blunt: "If you bootstrap to 100 million ARR AI business, anyone would call that a success. But I don't think it would be considered a success if you raised 100 million."

Thinking

1. Bad news can force an opportunity. Losing the offer nearly sent him into depression, but it also gave him the time and the push to build his own product. In his own words, losing that job might have been the best thing that ever happened to him.

2. When a new technology appears, building beats researching. No one had used a "chat with your documents" product, so user interviews couldn't give answers. He trusted his own judgment, built it and watched how the market reacted.

3. Speed is the biggest early advantage. He shipped a first version in six weeks and put up a pricing page on day one. He admits his reason for rushing "was maybe not right," but the urgency itself was right.

4. Charging from the start lets you skip fundraising. Someone paid 30 minutes after the pricing page went up. Revenue supported the company from day one, so he never needed to raise money to survive.

5. The product has to change with the market. From "chat with your PDF" to an AI customer service platform, Chatbase kept adjusting. In a fast-moving industry, being willing to change your mind matters more than sticking to the original plan.

6. Bootstrapping makes it easy to be too conservative. He thinks he hired too slowly and spent too cautiously early on. Once you have enough profit and conviction, you need to be willing to invest more.

Action

1. Play with new tools and watch for what just became possible. When new models and APIs come out, ask what they can do that wasn't possible before. That's often where the opportunities are.

2. Ship a usable first version as fast as you can. Don't wait for perfect features. Build a version that shows the core value, record a demo and post it.

3. Put up a pricing page on day one. Even with a few dozen users, let the people who want to pay do so. Whether anyone pays is the most honest test of demand.

4. Share your progress publicly. Post demos and numbers on X and similar platforms. Yasser had 16 followers at launch, but one good demo tweet can reach hundreds of thousands of people.

5. Hire people you trust first. When you need help, start with friends or classmates you already know and trust; they'll shape the company's early culture.

6. Check regularly whether you're being too conservative. If your business is steadily profitable, ask yourself: are there investments you've skipped just because you didn't want to spend the money?