Built This Week/AI Tools

AI Tools

Supabase

Supabase is the default backend under almost every internal tool we vibe code: auth, the database and edge functions, simple enough that Sam, a non-developer, wires it up with AI help.

By Jordan Metzner & Sam Nadler · From our episodes · Updated October 2026

Watch: Episode 4, Why Our Recruiters Stopped Reading Resumes And Let AI Take Over Everything from 08:25

Our verdict on Supabase

Supabase is the backend we keep coming back to. Seven of the apps we have shown on the show store their data in it, from Ryz Score in July 2025 to the Offsiteio estimator in April 2026. We have never spent an episode reviewing it, because it mostly just works and lets the AI coding tool do the rest.

What we built with Supabase

  • Ryz Score (July 2025): Jordan's AI resume scorer. Supabase handles authentication, authorization and storage, and edge functions check for new candidates and send each one to ChatGPT with a scoring matrix. Scores flow back into our applicant tracking system.
  • Mate (July 2025): our internal company directory, scaffolded in Bolt and finished in Cursor over roughly a weekend. Supabase is the database of record, plus search and authentication.
  • Daily Check-in (August 2025): Sam's digital stand-up. Submissions live in Supabase, an edge function summarizes them with AI, blockers first, and a GitHub cron job posts the summary to Slack each morning. Sam put about two hours into it in total.
  • AI compliance training and certification platform (November 2025): a Supabase back end with a Vite and React front end, and scenario videos made with Google Veo.
  • ScreenEval (January 2026): Sam's recruiter coaching app, built in about six hours with Claude Code and Supabase.
  • Hollywood transformation trainer (January 2026): when Claude Cowork asked Sam to pick a stack live on air, he chose React and Supabase.
  • Offsiteio estimator (April 2026): started in Claude Code, moved to OpenAI Codex, with Supabase as the backend. Estimate turnaround fell from about a week and a half to three or four days.

How we use Supabase

  1. Build the front end with an AI tool, then connect Supabase. Jordan said Bolt integrates with Supabase directly, so you can scaffold there and move the code to Cursor to finish (Episode 5, at 09:29).
  2. Let Supabase own auth and data. Ryz Score and Mate both use Supabase authentication, and every record lives in the Supabase database.
  3. Put the AI calls in edge functions. Ryz Score's edge functions watch for new candidates and call ChatGPT. Daily Check-in's summarize every team's check-in with a prompt that puts blockers first.
  4. Schedule jobs outside the app. Sam used a GitHub cron job to post the morning summary a little after 10am Central.
  5. Tell your coding agent the stack up front. In the Cowork demo, the agent proposed trade-offs and Sam answered "React and Supabase" before it started building.
  6. Deploy the front end separately. ScreenEval and the estimator run as React apps on AWS Amplify and GitHub, with Supabase behind them.
And then we use Supabase edge functions to check whether or not we have a new candidate in the pipeline. And then from those from those edge functions, we then send that data into ChatGPT with a scoring matrix as well as the job description and their resume and full application.

— Jordan Metzner, Episode 4

Where Supabase falls short

Honestly, we have not hit a wall with Supabase on the show, and we have not covered its pricing, limits or scaling. Our apps are internal tools for one team, so treat this as a builder's view, not a load test.

The one gap we worked around is scheduling. Sam ran Daily Check-in's morning job from GitHub, not from Supabase. And in Episode 8, Jordan said his Public API trading bot would have needed a proper backend, probably Supabase, with a React front end and cron jobs, if he had more time. Supabase is only the data layer, so plan for hosting and scheduling too.

Built it in Bolt, and, yeah, we use Supabase for the database and for the edge functions, and I used GitHub for the cron job.

— Sam Nadler, Episode 7

Episodes featuring Supabase

  • Episode 4: Ryz Score's auth, storage and edge functions.
  • Episode 5: Mate, from Bolt to Cursor on Supabase.
  • Episode 7: Sam's first edge functions and cron job.
  • Episode 8: what Jordan would add for a real trading bot backend.
  • Episode 20: the compliance platform's back end.
  • Episode 29: ScreenEval and the Cowork live build.
  • Episode 38: the Offsiteio estimator.
  • Episode 42: Karim Malhas names Supabase as a model for how Breeze organizes projects, teams and monitoring.

FAQ

What is Supabase used for?

In our builds it handles authentication, authorization, the database of record and edge functions. Ryz Score uses edge functions to spot new candidates and send them to ChatGPT for scoring, and Daily Check-in uses them to summarize stand-ups with AI.

Can a non-developer use Supabase?

Yes. Sam has never written a line of code and used Supabase for Daily Check-in, ScreenEval and the Offsiteio estimator. He says he finds it very easy to understand, and he learned edge functions and cron jobs with AI help.

Does Supabase work with Bolt, Cursor and Claude Code?

Yes. Jordan said Bolt integrates with Supabase directly, so he scaffolded Mate in Bolt and finished it in Cursor. Sam's later apps pair Supabase with Claude Code and OpenAI Codex.

Where do you host apps that use Supabase?

Supabase holds the data, and the front end is a React app. ScreenEval and the Offsiteio estimator deploy through AWS Amplify and GitHub.

How much does Supabase cost?

We have not discussed Supabase pricing on the show. The compliance platform was described as costing essentially nothing beyond development time, but that was about the whole build, not a Supabase price.