In partnership with
Happy Thursday 🎉 Another week, another batch of builders shipping faster than entire product teams. In this issue, we’ve got the crawler startup that fell into a $14.5M round (literally), an AI-built reading app clearing $60K/month, and a dev-vs-AI challenge where 5 prompts went head-to-head with senior-level code. Plus: open-source growth, AI agents, and real tactics you can steal.
.
In this issue:
A chair-fall and a cold email unlock $14.5M for Firecrawl 🤑
A 1-page AI app hits $60K/month—built entirely with prompts ✍️
Grammarly ships 8 new AI agents and a full-stack writing platform 🤖
Papermark grows to $900K ARR w/ a tweet & open-source strategy 📈
Can AI outbuild a senior dev? 👀
Training cutting edge AI? Unlock the data advantage today.
If you’re building or fine-tuning generative AI models, this guide is your shortcut to smarter AI model training. Learn how Shutterstock’s multimodal datasets—grounded in measurable user behavior—can help you reduce legal risk, boost creative diversity, and improve model reliability.
Inside, you’ll uncover why scraped data and aesthetic proxies often fall short—and how to use clustering methods and semantic evaluation to refine your dataset and your outputs. Designed for AI leaders, product teams, and ML engineers, this guide walks through how to identify refinement-worthy data, align with generative preferences, and validate progress with confidence.
Whether you're optimizing alignment, output quality, or time-to-value, this playbook gives you a data advantage. Download the guide and train your models with data built for performance.
News
The crawler startup that fell into a $14.5M round
TL;DR: Firecrawl, the AI-powered web data crawler for developers, secures $14.5M in Series A funding, powering over 1 billion requests and nearing 50k GitHub stars.
Firecrawl powers agentic browsing with an open-source crawler and commercial API—already used by 350K+ devs and trusted by Shopify, Replit, and top hedge funds.
Their $14.5M Series A came together after a literal chair-falling moment with Nexus and a gutsy cold email that pulled in Shopify CEO Tobi Lütke.
They went viral trying to hire an AI agent; that failed, but now there's a $1M bounty and a new opening for an AI chief of staff.
Firecrawl wants to rewrite the rules of AI crawling—by helping publishers get paid when their content trains or powers AI.
As AI dev tools explode, Firecrawl's success democratizes web data access for coders and entrepreneurs, but raises questions about ethical scraping and data privacy in an era where every startup wants to build on vast internet corpora—potentially reshaping SaaS for AI builders.
Idea Generation
TL;DR: Mikey walks through building a distraction-free reading app complete with Stripe subscriptions and AdSense ads using Replit AI and zero written code. The app, which now pulls in $60K/month, shows how AI lets solo founders ship focused, profitable tools by prompting, not programming.
The app solves one clear problem: clean, distraction-free reading with offline access, dark mode, and customization.
Stripe integration powers a $19.99/year premium tier; Google AdSense monetizes free users without cluttering the UI.
Every fix, from broken login flows to payment bugs, was handled via iterative prompting.
Success came down to precision prompts, lean feature scope, and ruthless UX focus.
You don’t need a full-stack team or deep tech to build a money-making SaaS in 2025. With AI, clarity of problem, UX detail, and smart monetization matter more than code.
Productivity
Grammarly launches 8 agents for research, grading, and reader insight
TL;DR: Grammarly just launched eight specialized AI agents and a new AI-native writing surface called “docs” aimed at students and professionals. These agents help users do everything from finding credible sources to predicting how a reader will react—all while keeping the user in control.
Reader Reactions and AI Grader give real-time feedback on clarity, tone, and alignment with audience or rubric so users can course-correct before hitting send.
Citation Finder and Expert Review surface credible sources and subject-specific feedback, turning AI into a research assistant that actually understands context.
Grammarly’s new “docs” is an AI-native writing surface where agents work alongside you.
For students and pros alike, Grammarly’s agents promote responsible AI use, boosting skill-building, clarity, and confidence without outsourcing the thinking.
This is a major pivot by Grammarly. They’re moving from grammar suggestions to full-stack writing assistance, tailored to context, goals, and audience. These tools are available to both Free and Pro Grammarly users, although plagiarism checks and AI detection are limited to Pro accounts. More features will launch later this year.
Built for Managers, Not Engineers
AI isn’t just for developers. The AI Report gives business leaders daily, practical insights you can apply to ops, sales, marketing, and strategy.
No tech jargon. No wasted time. Just actionable tools to help you lead smarter.
Start where it counts.
Founder Story
Hittting $900K with open source and a single tweet
TL;DR: Papermark, an open-source Docsend alternative, grew from a weekend project to $75K MRR in under 18 months. Co-founders Mark and Julia turned a viral tweet into a thriving business using an open-core model, global community contributions, and a lean cost structure.
Started as a weekend MVP, Papermark now generates $900K ARR through paid hosting and enterprise licenses layered on a free self-hosted core.
One tweet sparked 100K views and kicked off the product’s early traction—underscoring the power of building in public.
Community-driven development accelerated product velocity and trust; open code invited contributions, audits, and global reach.
Their model turns transparency into a competitive edge, especially in trust-sensitive categories like document sharing and virtual data rooms.
Papermark’s transparency, speed of iteration, and community engagement gave them a defensible edge over bloated incumbents. In an era of niche AI tools and agile dev, open source is increasingly the smarter go-to-market.
Tutorial/Framework
Can AI outbuild a senior iOS dev—with just 5 prompts?
TL;DR: A vibe coder and a senior iOS engineer went head-to-head in a no-code app dev challenge: clone Granola, a $250M AI meeting app, using only 5 AI prompts each. Vishall used Claude Code and Swift in Xcode. Riley used Vive Code with Opus 4.1. No manual edits allowed. Both apps shipped—with real audio recording, transcription, summaries, and calendar sync.
Riley relied on Whisper for accurate transcription and built clean UI with strong folder management, but delayed calendar integration.
Vishall used native Apple frameworks, nailed calendar sync, but hit friction with build errors and folder implementation.
Every bug fix, feature tweak, and UI update had to be solved via AI prompt—no direct code edits, only strategic prompt engineering.
The experiment shows that prompt design is now a core dev skill, shaping how features are prioritized, tested, and refined in AI-led builds.
AI can build feature-rich apps with minimal human input—but prompt strategy, platform familiarity, and debugging instincts still make or break the end result.
Want to sponsor this newsletter? This newsletter goes out to thousands of creators, builders, and early-stage founders every week. If you’d like to get your product or brand in front of them, you can book a sponsorship slot here
