Restaurant Vibes, Coding
Restaurant Vibecoding
Audio by Eleven Labs can be a little wonky
Apologies for the resend- I had a significant copy and paste mistake—oops!
TL;DR
Vibecoding = using AI to build software without coding knowledge
Restaurant tech companies build for the generic 80% of needs, missing the crucial 20% that makes each restaurant unique
I built custom scheduling and sales analysis tools in days that supplement core tech tools but, with some help from professionals, could replace them
Operators shouldn't wait for tech companies to catch up - we can solve our own problems now
This approach could create new high-margin revenue streams for restaurants sharing tools with each other
Links to tools at the end
Vibecoding is having a moment. The term was coined by Andrej Karpathy, an original OpenAI founder, and he describes it as "where you fully give in to the vibes, embrace exponentials, and forget that the code even exists." Another Karpathy quote, "English is the hottest new coding language," sums the idea up well. Vibecoding means using no-code tools to build apps and other solutions by just describing what you want to build. While Karpathy is talking about how this impacts developers and those who write code, the truly revolutionary aspect is how it unlocks capabilities for people like me. Forgetting the code exists is easy when you have no idea what it means or is doing.
I've felt over the past year that technology has finally caught up to where restaurants need it to be. I need to be able to talk to technology like another person and now I can—with one major (and frustrating) caveat: I can't use any of my restaurant-specific technologies (Toast, OpenTable, etc.) in this way. That's where restaurant vibecoding comes in. Before we get any deeper, let's talk about how I’ve been using it in our restaurants.
From Generic to Custom: Solving Real Restaurant Problems
We use Sling for scheduling in our restaurants. Before that we used Homebase and have also used 7 Shifts, Schedule Fly, and Hot Schedules in previous restaurants. I honestly couldn't tell you what differentiates these products from one another. They all have drag and drop scheduling, availability, schedule switching, messaging, and a few other features. None of them understand any of the nuances of scheduling. I can't tell them to avoid "clopens" (late shifts into opening shifts that don't allow for rest). I can't easily, if at all, give them pars for shifts and availability and have them compose a schedule draft to review. And I certainly can't chat with them naturally about availability, business volume, or team performance.
In a few days, I was able to build a scheduling tool that I could customize for nuances like minimum rest time between shifts or no weekend doubles. I provide it with pars (staffing levels by shift and day of week), shift start times, and our team's availability (all using simple Google forms), and it creates the schedule based on the pars, team availability, and my custom set of rules. It even has a chatbot I can ask questions like "I need extra support this Friday, who's available and what's the best option given the current schedule?" It plots the schedule to about 80% accuracy before someone has to come in to make the final adjustments. Even those last tweaks could probably be built in, but the time it would take to implement them is more than the few minutes needed for manual edits.
Sales analysis is another area where I frequently use vibecoding tools like Lovable, Replit, Bolt, or v0.dev. I've always felt solving problems in restaurants is about finding a needle in a haystack. Traditional tech tools like POS systems basically give you the haystack and the ability to break it into a few distinct piles to narrow your search. They're not good at analyzing across multiple parameters. For example, if our sales are down, is it because of fewer guests or lower averages per guest? If it's lower averages per guest, is it a particular daypart or sales category driving the trend? If liquor sales are the primary culprit, is there a specific area of the restaurant where the trend is strongest? Can I find all of this information in Toast? Yes. Can I do it as efficiently as restaurant operations demand? No! I now process four sales reports every week to get this information in seconds rather than hours.
The Wider Impact: Transforming the Restaurant Industry
The implications of vibecoding for the restaurant industry extend far beyond individual operational improvements. As restaurant owners and managers gain the ability to create custom solutions without coding knowledge, we're likely to see a democratization of restaurant technology that could fundamentally change how the industry operates.
Every restaurant has its own operational DNA—unique service styles, menu engineering approaches, staffing patterns, and customer expectations. Traditional restaurant tech has forced operators to adapt their processes to fit standardized software, often resulting in operational compromises.
With vibecoding, restaurant-specific workflows that were previously impossible to automate can now be custom-built by the very people who understand them best: the operators themselves. This shift enables restaurants to preserve what makes them special while still benefiting from technological efficiency.
The 80/20 Problem Solved
The restaurant tech industry has long focused on building for the 80% use case—the common denominators across all restaurants. This approach makes business sense for SaaS companies seeking scale, but it leaves restaurants struggling with the 20% of operations that truly differentiate them from competitors. This 20% isn't trivial—it's often the heart of a restaurant's identity and competitive advantage. Traditional tech companies simply cannot build economically viable products that address the infinite variations of that crucial 20%.
Vibecoding fundamentally changes this equation. Tools like Lovable, Bolt, Replit, and v0.dev enable operators to fill these critical gaps themselves, without waiting for tech companies to build solutions that may never come. Beyond the tools I've built so far, here are some other possibilities:
Inventory Management: Create custom tools that account for your specific prep cycles, waste patterns, and ingredient cross-utilization
Menu Engineering: Build analytics that track menu item performance based on your unique profitability metrics, not generic industry standards
Staff Training: Develop interactive training modules that reflect your restaurant's service philosophy and specific guest interactions
Guest Experience: Create custom CRM tools that track preferences specific to your concept and clientele
Kitchen Operations: Design workflow tools that match your kitchen layout and cooking methods
The Industry Crossroads
Restaurant tech companies must decide: will they embrace this shift by creating platforms that enable operator customization through vibecoding interfaces, or will they continue forcing restaurants to adapt to their one-size-fits-most approach? I've spoken to a number of founders and senior executives from restaurant tech companies over the past few weeks, and I do truly believe there is genuine interest in using these tools to democratize restaurant tech. The problem lies in the classic innovator's dilemma. Why disrupt my successful core business for an innovation that might cause short term pain but has the potential for long term gain?
While restaurant tech companies could potentially build platforms that incorporate vibecoding capabilities—allowing operators to customize within their ecosystems—the question remains whether they'll move fast enough to prevent operators from simply bypassing them entirely. The most forward-thinking tech companies in this space will likely create platforms where core functionalities (POS, reservation systems, inventory) can be extended through operator-designed customizations. However, the economic incentives to maintain control will slow this transition.
The Path Forward: Don't Wait
In the meantime, I don't think we, as operators, should wait. By embracing tools like Bolt and Lovable today, we can start building the custom 20% that makes our restaurants special, while still using traditional systems for core functions. Doing so could also help restaurants help each other. I imagine there are other restaurants that would see value in the sales analysis and scheduling tools I've described. Creating digital tools for ourselves and other operators could mean high-margin revenue streams for our painfully low-margin industry.
The real breakthrough will be unlocking creativity and entirely new markets for restaurants. I used v0.dev to build a prototype of an adtech tool that allows brands to prospect and advertise through restaurants (working title: Focal). In conjunction with the classification underlying Reggi, The Restaurant Genome project, the vision is to create an advertising platform that helps restaurants monetize around attention and community.
This vibe shift doesn't necessarily mean abandoning established tech platforms entirely—at least not yet. Instead, it creates a hybrid approach where operators use standard systems for common functions while building custom tools to address their unique operational challenges.
The true revolution lies in the power to solve problems independently, without waiting for tech companies to recognize and address your specific needs. In an industry where margins are thin and every operational efficiency matters, this independence could be transformative.
Tools and Links:
Most of these have a very limited free version and then paid plans start at $20 a month. You can build most small things within that tier but larger projects might drive you closer to $100 per month.
Lovable: My Current Vibe Code Tool of Choice
And a few others I’ve use and really like:
Claude by Anthropic: My favorite tool for back and forth in thinking through this stuff
Great read. The concept of vibecoding and the fundamental 80/20 shift (and the openness of the conversation of restaurant operations now) - is super cool.
One line that stood out: "Doing so could also help restaurants help each other." That idea I think could get lost in the excitement around creative Guest XP tools, scheduling software, kitchen workflow - but I wonder if it might be a key in long-term sustainability for the neighbourhood spot.
As an independent operator, you’re often stuck with rigid, inherited systems that weren’t built for you but are still being updated for everybody. And really these are pretty good all considering. But dishwashers break, staff don't show up - shit is constantly hitting the fan. Time goes by, and it's hard to to zoom out to assess the real decisions at any given time. What's our identity? What captivates Our customers? With media platforms giving you a thousand ways to express yourself as a business, these questions go beyond the four walls. To me, generative ai tools, and this direction, shift that dynamic entirely.
The idea of restaurants helping each other reframes what we even need from our systems.
Is this the return of the restaurant association? Can we share the burden of developing smarter, more adaptive tools that are both data driven, deeply personal and cool af? What do these shared partnerships look like? Is sharing the way forward? Anyway - great writing, I find this stuff super cool.
Killer post. Love it!