✦ CESXY
Submit a buildSign in

Small projects made with AI

Built on a whim.
Proven in the wild.

CESXY publishes small projects made with AI, after an editor has opened them and tried them. Each listing says what worked, what it asks of you and what to watch for.

Discover projects →Submit a build →
CESXY Field notesField notesHow to judge an AI-made app in five minutes before you trust itA five-minute check for any small AI-made app: who made it, what it asks for, login and payment warning signs, and a test of the core action.CESXY Editors · 4 minCESXY Field notesField notesWhy we curate by reason, not ranking: how CESXY collections workCESXY Collections group projects for a stated reason instead of by popularity. Here is how a collection is put together and kept current.CESXY Editors · 3 min

Editor's desk

From the editor's desk

Short notes on what we tested this week and what we learned from it.

All from the desk →
An editor’s desk with a phone prototype, design notes, and a small keyboard

How we review: what "tested" means on CESXY, and what it does not

What an editor checks before a project is listed, what each badge means, and why a badge is a use note rather than a certification.

Read the guide →

Read next

CESXY Field notes
What is CESXY?CESXY publishes small projects made with AI after a person has opened them and tried them. Here is how it works and who it is for.
CESXY Field notes
How to submit your AI-made project to CESXY and prepare it for publicationWhat qualifies for CESXY, how to write a description that matches your project, what happens after you submit, and common reasons a build is not listed.
CESXY Field notes
Launch checklist for AI-built Web apps: twelve checks before strangers try itTwelve checks to run before strangers try your AI-built web app, from broken links and phone layout to privacy notes and clear pricing.

Hardware prototyping

Proto

Hands-on guides to building small hardware with ESP32 boards, sensors and a little code.

All from Proto →
Xiaozhi on ESP32: A Practical Guide to Building Your Own AI Voice Assistant article cover

Xiaozhi on ESP32: a practical guide to building your own AI voice assistant

Xiaozhi is open-source firmware that turns a low-cost ESP32-S3 board into a talking assistant, and it is one of the easiest ways for a beginner to put a large language model inside a small device.

Read the guide →

Read next

CESXY Field notes
Running your own server for an ESP32 voice assistant with Xiaozhi-esp32-serverThe open-source xiaozhi-esp32-server lets you choose the speech, language and voice models behind a Xiaozhi device, keep data on your own machine and add tools, memory and voice recognition.
CESXY Field notes
ElatoAI: how an ESP32 turns a toy into a real-time voice companionElatoAI is an open-source stack that connects a small ESP32-S3 board to real-time voice models, so a plush toy or desk gadget can hold a spoken conversation without a phone in the loop.
CESXY Field notes
A language model on a microcontroller: what the esp32-AI project shows and what it does notThe open-source esp32-ai project runs a 28.9 million parameter language model directly on an ESP32-S3 by keeping most of it in flash memory, and its honest limits teach as much as its results.
CESXY Field notes
Clawdmeter and desk dashboards: putting your AI coding usage on a small ESP32 screenClawdmeter is an open-source ESP32 desk display that shows how much of your AI coding plan you have used, and it is a clear example of how a host computer, Bluetooth and a small screen can work together.

Working right now

Projects an editor has checked recently and found doing what they say.

Browse all →
Travel ImmersiveTravel planning

Travel Immersive is one of those apps that makes a clear case for spatial computing: it turns a catalog of faraway places into something you can look around inside. On Apple Vision Pro, its strong collection, polished feature set and thoughtful social touches make it an appealing way to spend time somewhere new—even when you only have a few minutes. Now the Vision Immersive Team is giving away complimentary 1-month premium access to early evaluators! Limited Quantity: Only 50 codes available for Apple Vision Pro & 50 codes for Meta Quest

Vision Immersive Team

Browse by category

All categories →
Productivity 0→AI Tools 0→Entertainment 0→Games 0→Education 0→Lifestyle 0→Health & Fitness 0→Finance 0→Travel 1→Creativity 0→

For builders

Made something small?
Let real people use it.

Send us a public project you have the right to share. An editor will use it, check the basics and write an honest note. You can also start from a wish someone else posted.

Submit a build →
  1. Wish it.Someone posts a need on the Wishes board, or you notice one in your own life.
  2. Shape it.Cut the idea down to the one action a person needs to complete.
  3. Build it.Make that action work at a public address, with AI doing as much of the work as you like.
  4. Submit it.Send it to CESXY so an editor can try it and decide whether to list it.

How labels work

Tested isn't certified.
It's a use note.

Badges are earned by testing and are never sold. Creators do not see our verdict before it is published. A label records what an editor saw on the day; it is not a guarantee.

How we review →
An archival CESXY dossier with notes, a folded map, and a green line

The short version

What sits behind
A CESXY label.

Before you trust a listing, you should know who wrote it and on what terms. Here are the four facts that matter.

  1. Who tests.CESXY editors, who open each project and use it the way a normal visitor would.
  2. What we check.That the address loads and the core action works, plus login, payment, mobile use and obvious limits.
  3. What it costs.Free review and listing cost nothing; paid deep testing is a separate service and never buys a badge.
  4. Corrections.If we got a fact wrong, email us and we will check it and fix the listing.
Read how we review →

Get the CESXY briefing

New reviews and tools worth your time, about once a week. Unsubscribe anytime.