While our SEO and content strategies provided steady growth, we also experienced several breakthrough moments that rapidly accelerated user acquisition. These moments weren't accidental. They came from deliberately positioning ourselves at the intersection of emerging trends and real user needs, and most of them came from one source: the AI wave that began reshaping the internet in 2022.
This chapter is about how we rode that wave, and how we tried to do it without losing sight of what Glasp was for.
The Aurelius Effect: One Right Partnership
Before AI entered the picture, our first major growth acceleration came from a strategic YouTube sponsorship that returned far more than we expected.
By August 2021, we had gathered enough evidence about our target audience to consider a YouTube sponsorship. Rather than spreading a small budget across many creators, we concentrated our resources on one partnership.
After researching dozens of potential partners, we identified Aurelius, an Australian YouTuber with around 300,000 subscribers whose audience aligned perfectly with our target users: thoughtful knowledge workers interested in productivity and learning.
The sponsorship itself was straightforward, a brief mention of Glasp in one of his videos. What we didn't anticipate was the cascade it set off:
- Aurelius featured Glasp in a video, driving the initial wave of sign-ups
- Other creators who watched Aurelius began discovering and mentioning Glasp
- These creators made their own videos, posts, and articles about Glasp
- Their audiences discovered Glasp, and the most influential among them created even more content
This "influencer cascade" created a multiplier effect far beyond our initial investment. In the creator economy, reaching the right person can mean indirectly reaching their entire network of peer creators. When resources are limited, depth of partnership beats breadth.
DALLE-dle: An Early Experiment at the Edge of AI
In mid-2022, we began noticing early signals of the coming AI revolution. DALL-E had just been released, demonstrating impressive image generation. At the same time, Wordle had become a global phenomenon as a simple, engaging word game.
Rather than merely observing these trends, we asked: how can we combine them with Glasp's core functionality to create something unique?
The result was "DALL-E Wordle" (or "DALLE-dle"), a game where players guessed which famous quote had been used to generate an AI image. We used quotes from the highlights collected on Glasp, creating a natural connection to our core product.
The experiment was featured in PC Gamer, a major gaming publication, as an interesting new take on the Wordle format, driving significant traffic to Glasp. DALLE-dle wasn't our core product, but it created awareness and positioned us as builders in the AI space just as public interest was beginning to surge.
The ChatGPT Chrome Extension: Days, Not Weeks
When ChatGPT launched in November 2022, we immediately recognized its potential. Rather than viewing it as competition, we asked how we could build with it.
Within days of ChatGPT's release, we created one of the first Chrome extensions for ChatGPT, letting users access the AI assistant from any webpage. This simple tool addressed a real friction point for early ChatGPT users.
The timing mattered enormously. In its first weeks, our extension became one of the most installed ChatGPT-related Chrome extensions, quickly reaching 300,000 installations and bringing a massive new audience into our ecosystem.
This wasn't luck. It was the result of constantly monitoring technological trends and being prepared to act fast when an opening appeared. By being among the first to build useful tools around ChatGPT, we captured attention that would have been far harder to earn even weeks later, once the market was saturated.
YouTube Summary with ChatGPT: The Tool That Outgrew Us
Our biggest viral success combined two things we already had: YouTube highlighting (which predated ChatGPT) and our new AI integration experience.
YouTube contains an enormous amount of valuable knowledge, but extracting and retaining it is hard. We built the YouTube Summary with ChatGPT, a Chrome extension that could:
- Automatically extract the transcript from any YouTube video
- Run that transcript through an AI model to generate a concise summary
- Present the summary alongside the video for easy reference
- Let users save both the summary and their own highlights to Glasp
It addressed a clear pain point: the time it takes to extract knowledge from long-form video. Instead of watching a 30-minute video, users could read a summary in 30 seconds, then decide whether the full content deserved their time.
The response was immediate and overwhelming. Users shared screenshots across social media, generating organic promotion we could never have afforded to buy. YouTube creators made videos showing how it improved their research process. Newsletter writers included it in their roundups of essential AI tools. Business publications like Forbes covered it as an innovative use of AI.
Within a year, the YouTube Summary with ChatGPT had been installed by over 2 million users, more than double our core Glasp user base at the time. Our extension had become more popular than our product.
Rather than seeing that as a problem, we treated it as an opportunity. The YouTube Summary wasn't a standalone tool; it connected directly to users' Glasp accounts, introducing them to highlighting, note-taking, and knowledge sharing. Users who came for the AI summary often stayed for the rest. It became a natural upgrade path from casual AI tool user to committed Glasp user.
Digital Clones: Your Highlights, Talking Back
Building on that momentum, we developed Digital Clones: personalized AI assistants trained on a user's Glasp highlights and notes.
The concept was compelling. As you highlight articles, books, and videos in Glasp, you're effectively creating a dataset that reflects your interests, knowledge, and thinking patterns. Train an AI on that personal dataset and you get a digital extension of your mind, a "clone" that can discuss topics in a way that reflects your perspective.
Digital Clones solved several problems at once:
- It gave users a way to explore and interact with their collected knowledge
- It provided a compelling reason to keep highlighting and saving content
- It made our mission literal: knowledge that engages with others even beyond the individual
It wasn't as immediately viral as the YouTube Summary with ChatGPT, but it attracted attention from technology publications and futurists who saw it as a glimpse of where personal knowledge management was heading. And it created a powerful engagement loop: the more users highlighted, the more useful their clone became.
The Supporting Cast
Around these flagship features, we shipped a series of smaller AI experiments and systems:
- Idea Hatch used language models to find meaningful relationships between seemingly unrelated pieces of content in a user's collection, turning a passive archive into an active thinking partner.
- Content automation helped our two-person team operate like a much larger one. A single YouTube video could become a structured blog post, social snippets, and newsletter content, with humans reviewing and refining rather than drafting from scratch.
- The Kindle Personality Test analyzed a user's Kindle highlights to generate playful insights about their reading patterns and recommend books. Low development cost, naturally shareable results, and a clear connection back to reading and highlighting.
None of these were core products. All of them compounded attention, trust, and sign-ups.
What These Wins Had in Common
Looking across the Aurelius cascade, DALLE-dle, the ChatGPT extension, the YouTube Summary with ChatGPT, and Digital Clones, the same principles kept showing up:
1. Be first, even if imperfect
Our ChatGPT extension and YouTube Summary with ChatGPT weren't the most feature-rich options that eventually emerged, but they were among the first available when interest peaked. The window for a new trend is measured in days, not weeks. We prioritized speed above polish, while holding a minimum quality bar, and iterated after capturing attention.
2. Solve real problems tied to your core value
The YouTube Summary succeeded because it killed a real pain (hours of video, minutes of time), not because it showed off AI. And every feature we shipped connected back to knowledge capture and sharing. Trend-chasing for its own sake produces traffic; trend-chasing in service of your mission produces users.
3. Build distribution into the product
Our most viral features had sharing built in. The YouTube Summary's subtle branding appeared in every screenshot users posted. Digital Clones made people want to show others their personal AI. The product did its own marketing.
4. Monitor early signals
None of these opportunities appeared without warning. We noticed rising interest in AI image generation before DALLE-dle, and the early excitement around ChatGPT before our extension. Paying attention is a strategy.
5. Enhance the core, never replace it
AI features made highlighting more valuable; they didn't substitute for it. Digital Clones got better the more you highlighted. Idea Hatch got better the more you saved. Each feature deepened engagement with the core product instead of diverting from it.
The Balancing Act: Trends vs. Long-Term Vision
These viral moments accelerated growth dramatically, but they carried a risk: becoming known for trendy features rather than our mission of knowledge sharing.
We navigated that tension by insisting that new features, no matter how trendy, served the long-term vision. The YouTube Summary with ChatGPT wasn't an AI gimmick. It genuinely helped people extract knowledge from video, which is exactly what Glasp exists to do. We passed on many viral opportunities that didn't fit, even when they might have driven short-term growth.
In the end, these moments weren't separate from our sustainable growth strategy. They were accelerators layered on top of it. By connecting new technologies to an enduring mission, the spikes settled into sustained engagement rather than temporary attention.