// Blog

Build logs, insights,
and lessons from the trenches.

How we ship production software in days, working from one stack and an AI-native operating system. Technical deep dives, case studies, and the real numbers behind AI-native development.

recipeastroconvextech-stackengineering

Convex + Astro: The Real-Time Backend Pattern We Use

How we wire Convex into an Astro site for real-time data without giving up Astro's static-by-default speed. The recipe, plus the gotcha that breaks first attempts.

Read more →
recipeaisaasagentsmastrafounders

How to Add AI to Your SaaS Product: A Practical Guide

A practical guide to adding AI features to an existing SaaS: what to build, what it actually costs, what to skip, and the production traps to plan for upfront.

Read more →
startupfoundersmvpspeedproduct-market-fit

Speed Is the Only Moat for Early-Stage Startups

42% of startups fail building the wrong thing. Only 3% hit $1M ARR inside a year. The winners ship faster, learn faster, and re-decide weekly. Here is how.

Read more →
vibe-codingai-toolssecurityfoundersproduction

Vibe Coding Is Fine. Vibe Shipping Is Not.

Vibe coding produces great prototypes. Shipping them to production without review has caused data breaches, app store bans, and deleted databases.

Read more →
mvpcost-analysissaasfoundersstartup

What a $5K, $12K, and $50K MVP Look Like in 2026

A side-by-side comparison of MVP budgets in 2026: $5K validates demand, $12K builds the production moat, $50K adds enterprise features. Pick by stage.

Read more →
founderscofounderstartuphiringmvp

You Don't Need a Technical Cofounder in 2026

Stop searching for a technical cofounder. Developer shortage, equity fights, and hiring timelines make it the wrong move for most founders. Here is what works.

Read more →
mvpstartupsaascost-analysisfounders

The Real Cost of Building a SaaS MVP in 2026

What it actually costs to build a SaaS MVP in 2026. Side-by-side: hiring, freelancers, agencies, no-code, AI builders, and AI-native development. With numbers.

Read more →
ai-toolsmvpvibe-codingfoundersproduction

The 70% Problem: Why AI-Built MVPs Aren't Production-Ready

AI builders like Lovable and Bolt deliver ~70% of a working app. The remaining 30% costs $5K-$20K in engineering. Here is what breaks and how to plan for it.

Read more →
recipetech-stacksaasmvpengineeringnext-js

The Stack We Use for Every SaaS MVP (and Why)

Our opinionated SaaS MVP stack: Next.js, Convex, TypeScript, Tailwind, Mastra, Python, Vercel. Why we chose each piece, what we ruled out, what we'd revisit.

Read more →