# ChatGPT Sites examples: 40 real sites from the first three months · Stacktree Source: https://stacktr.ee/blog/chatgpt-sites-examples [Skip to content](#main) [Stacktree](/)[Developers](/developers)[Agents](/agents)[Docs](/docs)[Use cases](/use-cases)[Pricing](/pricing)[Blog](/blog)[Dashboard](https://app.stacktr.ee)[Sign in →](https://app.stacktr.ee) By [ Steve Smith ](/about) · Founder, Stacktree · Last updated September 11, 2026 blog · compiled 11 September 2026 # Forty ChatGPT Sites people actually built. OpenAI says five million Sites have been created in three months. Two threads on 11 September 2026 let you see a sample: OpenAI featured six in its announcement, and Kath Korevec's last call for an awesome list drew forty-seven replies from people showing their own. We collected every site with a working link, checked each one loads, and sorted them. The pattern is clear, and so is what is missing from it. [ Get started free ](https://app.stacktr.ee/?join=1&from=seo_blog_chatgpt_sites_examples) install npx stacktree-install Copy No card · 3 pages free · about a minute ## What do ChatGPT Sites examples look like? Mostly games, learning tools, small personal utilities and creative experiments, built by individuals and published publicly at a chatgpt.site subdomain that carries the maker's handle, with a handful on custom domains. The best ones do one job and hide the chat they came from. What the sample does not contain is client work: reports, proposals and audits, which need a private link a reader can open without a ChatGPT account. ## On this page - 01 [ Games and toys ](#games) - 02 [ Learning and reference ](#learning) - 03 [ Tools people use every day ](#tools) - 04 [ Creative and generative ](#creative) - 05 [ Personal sites and business fronts on custom domains ](#domains) - 06 [ What the forty have in common ](#pattern) - 07 [ What is not on the list ](#missing) - 08 [ Submitting your own ](#submit) ## Games and toys The largest category by a distance. Many were one-shotted; several are multiplayer. - [Pocket Grove](https://pocket-grove.hedgehoglw.chatgpt.site/). Cozy pixel-art pet game. Featured by OpenAI in the announcement thread. - [World Shiritori](https://world-shiritori.iori-2.chatgpt.site/). Shared Japanese word-chain game, one chain for everyone. Featured by OpenAI. - [Calculatron: Maximum Drama](https://calculatron-maximum-drama.kendonino.chatgpt.site/). A calculator that turns arithmetic into theater. Featured by OpenAI. - [Rill](https://rill.wojtekmaj.chatgpt.site/). Relaxing nature game. The whole prompt: "Build me a relaxing game, entirely freestyle, nature related maybe?" - [ForestRide](https://forestride.aracos.chatgpt.site/). A slow drive through a 3D forest. - [Jelly Hockey](https://jellyhockey.logicleaptech.chatgpt.site/). Soft-body hockey. One of three games the same person shipped in a week. - [Tidetown](https://tidetown.logicleaptech.chatgpt.site/). A small living town, from the same week. - [Morph](https://morph-creature-lab.logicleaptech.chatgpt.site/). A jelly creature studio, the third. - [Harvey, a game](https://harvey-a-game.goobis.chatgpt.site/). A short narrative game. - [Crazy Arcade](https://jerry-crazy-arcade.jerryjhlee.chatgpt.site/). A browser remake of the bomb-arena classic, shared as "try it out, it's free". - [Nuke Zone 3D](https://nuke-zone-3d-game.kocahmet1.chatgpt.site/). A multi-floor 3D shooter that climbs a city complex to the roof. - [Starlight Goose Garden](https://starlight-goose-garden.maplewashere.chatgpt.site/). A garden game, with a sibling called Jelly Break by the same maker. - [The Bureau of Tiny Disasters](https://tiny-disaster-bureau.arashi.chatgpt.site/). "A silly website to goof around and disconnect and recharge." - Coleslaw Classic. A browser MMO with dungeons, 17 skills and a mountain boss, built with Codex and hosted on Sites. Shared in the announcement replies without a public link. ## Learning and reference The most useful group, and the one with the most domain experts building for their own field. - [Anatomy Atlas](https://anatomy-atlas.brianp.chatgpt.site/). A 3D human anatomy atlas built by a physician, with WebMCP so an agent can generate custom videos and quizzes on it. "I would have LOVED having this as a medical student and resident." - [Anatomy Lab](https://external-anatomy-lab-maxwell.maxtoms21.chatgpt.site/). A second, independent anatomy reference. - [Original Transformer vs DeepSeek](https://transformer-architecture.petergostev.chatgpt.site/). An interactive walk through two transformer architectures side by side. - [Gospel Worlds](https://breakfast-by-the-sea.jackter.chatgpt.site/). The Gospels rendered as explorable 3D scenes. - [No Moat](https://no-moat.petergyang.chatgpt.site/). An essay-as-site on AI competition, submitted to the first call for the list. - [yanabana](https://yanabana.com/). A study of fashion history on the maker's own domain, called out by the Sites product lead as an example of what she likes seeing. - POKE-8. A 6502 assembly learning app and fantasy console with C64-style video memory, six example games, and a WebMCP hook so your AI can play along. Built "to try ChatGPT Work and Sites". - Inflation Lab. A teaching tool where students tamper with the inflation basket and watch the headline number move. Built to teach ChatGPT Work and Sites, tongue in cheek. ## Tools people use every day Small, personal, and doing one job. Several replaced a spreadsheet or a group chat. - [The Closet Index](https://the-closet-index.matijagrcic.chatgpt.site/). "Find the movie for the night you actually have." - [Tubi Dialer](https://tubi.griffinwbrown.com/). A dial-a-movie picker on the maker's own domain. "I guarantee you will find something for movie night." - [Sofreh](https://sofreh-recettes.emerika13.chatgpt.site/). A recipe site built entirely from a phone, using an MCP server to pull recipes out of Instagram, TikTok and YouTube videos. - [Bookshelf](https://booktower.elithrar.chatgpt.site/). A personal reading tower, the most-viewed submission in the thread. - [Reality Diff](https://reality-diff-world.bearhuddleston.chatgpt.site/). "A change log for the physical world": earthquakes, space weather and regional alerts in one feed. Featured by OpenAI. - [Taskelo](https://taskelo.io/). Task and project tracking on a custom domain, shared as "check what I did". - Padel league tracker. Weekly fixtures, points and a monthly ranking for a group of friends. Shared in Spanish with the advice to start with "something that you and yours will actually use". - [api.term.app](https://api.term.app/). "A forum for agents. It's just getting started." ## Creative and generative Where the newer models show off. Three of these were featured by OpenAI. - [BrushRove](https://model-paint-cartch.cartch.chatgpt.site/). Rotate and paint 3D models directly on their surfaces. Featured by OpenAI. - [Was the Light Real?](https://was-the-light-real.songsongmiaoling.chatgpt.site/). An interactive visual story about AI and connection. Featured by OpenAI. - [Oath & Relic](https://oath-relic-living-journal.bradjr.chatgpt.site/). A living tale that updates as a journal. - [Afterlight](https://afterlight-tidal-garden.bpow.chatgpt.site/). A shared world where every encounter leaves a trace and your environment depends on where you physically are. - [Digital Peony](https://peony.lewis-o.chatgpt.site/). "A peony with a twist." - [Cosimo Synth](https://cosimo-synth.andrew-martin-stern.chatgpt.site/). A full wavetable synthesizer in the browser. "Serum level," in the maker's words. - Astra Gallery. A 3D virtual gallery of 358 Codex Billboards, with AR so you can preview a work framed on your own wall. ## Personal sites and business fronts on custom domains Custom domains were the update people cheered loudest. These were already on them before the announcement. - [apollozac.com](https://apollozac.com/). A personal project page with an easter egg. "Usually don't have focus/incentive to actually ship a website, but ChatGPT Sites makes it really natural and fun." - [griffinwbrown.com](https://griffinwbrown.com/). A personal site, rebuilt with the newest model and shared as a before-and-after. - [ChatGPT Work guide](https://work-guide.openai.chatgpt.site/). OpenAI's own guide to ChatGPT Work, itself a Site, submitted to the list by a reader. ## What the forty have in common Each one does one job. The sharpest reply in the awesome-list thread was a rule rather than a link: "cool ones do one job, hide the chat chrome, and feel like a normal site." Every featured pick fits it. Pocket Grove is a pet. Reality Diff is a feed. The Closet Index picks a film. None of them try to be a platform. They are apps, not documents. Sites scaffolds a full application with a database and server logic behind every deployment, and the examples use that: multiplayer state in World Shiritori and Afterlight, a synthesizer engine, a live alert feed. This is the part ChatGPT Sites is genuinely good at, and it explains why the sample looks the way it does. People reach for a builder when they want to build. They are public, because they have to be. A site cannot be on a list unless anyone can open it. Two replies in the thread make the constraint visible: "Bummed I can't submit mine," and a filmmaker asking whether a Site "needs to be deployed publicly to qualify" because her storyboard tool is meant to be opened locally or shared with a team. Whatever people are building privately is invisible to a roundup like this one. The makers are individuals. A physician, a musician, an economics teacher, a filmmaker, several game developers, a class of twenty-five students whose instructor asked in the same thread whether the price could come down for them. Plus and Pro subscribers, mostly, building for themselves and their friends. The URL is part of the story. Every hosted address is site-name.handle.chatgpt.site, and the handle is yours. One reply to the announcement asked OpenAI to stop "doxing your email address as a part of the domain name." The makers who cared most moved to their own domains, which is why custom domains got the loudest cheer of the five updates. ## What is not on the list OpenAI's own inspiration list for Sites reads: "Launch dashboards to track and share results. Personal portfolios to help your work stand out. Onboarding hubs to help new teammates get up to speed. Games." Two of those four are internal. None of the forty examples above is a dashboard for a client, a proposal, an audit, a quarterly report or a design review. That is not because nobody makes those in ChatGPT. It is because those pages are private, their reader is outside your company, and as of the 11 September update the private option is an invitation the reader can only open by signing in with a ChatGPT account. A public gallery can only ever show you the public half. The replies under the five-million figure were skeptical for a related reason. "4.9 million of them are abandoned single-prompt wrappers with zero visitors," said one. "Websites are easy. Getting anyone to visit one is the hard part," said another. The number that would settle it is how many Sites have a reader who is not the maker, and for a client deliverable that is the only number that matters. Sites reports unique visitors and page views on public and workspace Sites. It does not tell you which named person opened the thing you sent them. If the page you are holding is the deliverable kind, the shape you need is different from everything on this list: [a private link the client opens with no account](/share-with-clients), a passcode or a company email-domain gate when the content warrants one, an expiry, your own domain, and a read on who opened it and how far they got. That is what Stacktree does, from any agent, over MCP or a single curl, and it is the half of the category a ChatGPT Sites gallery will never show. ## Submitting your own Kath Korevec (Codex, OpenAI) is collecting the list and has said it will be public when it is done. Reply to [her last-call thread](https://x.com/simpsoka/status/2098434201100136603) with a link. Her [feedback thread](https://x.com/simpsoka/status/2098244223728505322) from the same day is where the feature requests are landing: expiring links, a usage dashboard, built-in login for multiple users, exporting a Site with its git history. If you have a Site that is not public and want it seen, that thread is the place to say why. Fix it now You've read why the public link is a problem. Paste the artifact here and get a private one, no account, and a passcode if you want it. Paste HTMLDrop a fileno account · 24 h linkHTML to publishAdd a passwordPublish private link For the pages that cannot go on a list Reports, proposals, audits and client dashboards: private by default, opened with no account, on your own domain, with a read on who opened them. Three pages free, no card. [ Publish a private page → ](https://app.stacktr.ee/?from=blog_cta_chatgpt_sites_examples&join=1) Free to try: 3 pages, each live for 7 days. Own domain from $19/mo. FAQ ## Frequent questions What are some examples of ChatGPT Sites? + From OpenAI's own showcase on 11 September 2026: a pixel-art pet game (Pocket Grove), a shared Japanese word-chain game (World Shiritori), a 3D coloring app (BrushRove), a theatrical calculator, and a live change log of earthquakes and space weather (Reality Diff). From the community: a physician's 3D anatomy atlas, a wavetable synthesizer, a recipe site built from a phone, and several personal sites on custom domains. All forty are linked above. Is there a gallery or list of ChatGPT Sites? + OpenAI has not published one yet. Kath Korevec, who works on Codex at OpenAI, put out a "last call" on 11 September 2026 for submissions to an awesome list of ChatGPT Sites and confirmed it will be public when it is done. Until then, this page and the replies to her thread are the closest thing to a gallery. OpenAI also featured six sites in its announcement thread. Are ChatGPT Sites public? + They can be, and every site on a list like this one has to be, which is why the examples skew toward games, tools and art rather than work product. A Site is private to its owner until access is changed. Options are owner and admins, selected users or groups plus named external viewers (who must sign in with a ChatGPT account), anyone in the workspace, or anyone on the internet. What URL does a ChatGPT Site get? + A subdomain of chatgpt.site that includes your account handle, in the form site-name.handle.chatgpt.site, for example anatomy-atlas.brianp.chatgpt.site. One reply to the announcement objected to the handle being part of every address. Where custom domains are available you can connect an apex domain or subdomain you own by changing its DNS records. What do people build most with ChatGPT Sites? + Judging by forty submissions and OpenAI's own picks: games first, then learning and reference tools, then small personal utilities, then creative experiments, then personal sites. OpenAI's inspiration list adds dashboards, portfolios and onboarding hubs. Almost none of the public examples are client deliverables, because a deliverable is private and its reader usually has no ChatGPT account. Can I make a client report or proposal as a ChatGPT Site? + You can build it. Sharing it is the awkward part: the client either gets a public URL or an invitation they can only open by signing in with a ChatGPT account. There is no passcode, email-domain gate or expiry. For that case Stacktree publishes the same HTML to a private link a client opens with no account, gated if you want, on your own domain, with a read on who opened it. Keep reading ## Related guides - [ ChatGPT Sites explained Cost, access, domains, limits, and what is still undocumented. ](/blog/what-is-chatgpt-sites) - [ ChatGPT Sites private sharing Who can open an invited Site, and why they still sign in. ](/blog/chatgpt-sites-private-sharing) - [ ChatGPT Sites usage limits What the 50% warning means, and how not to spend quota on static pages. ](/blog/chatgpt-sites-usage-limits) - [ Stacktree examples The deliverable half: reports, proposals and audits as private pages. ](/examples) - [ Host AI-generated reports When the output is a document, not an app. ](/host-ai-reports) References ## Sources and further reading - [ ChatGPT Sites update: 5M sites, private sharing, custom domains (@ChatGPT) ↗ The 11 September 2026 announcement, with OpenAI's six featured sites and its inspiration list in the thread. ](https://x.com/ChatGPT/status/2098457920291946894) - [ Last call for an awesome list of ChatGPT Sites (Kath Korevec, OpenAI) ↗ The community thread most of the examples above come from; the list will be public when done. ](https://x.com/simpsoka/status/2098434201100136603) - [ ChatGPT Sites feedback thread (Kath Korevec, OpenAI) ↗ Feature requests from makers: expiry, usage dashboard, multi-user login, export with git history. ](https://x.com/simpsoka/status/2098244223728505322) - [ OpenAI Help Center: Creating and managing ChatGPT Sites ↗ Access options, the sign-in requirement for invited viewers, and public beta limits. ](https://help.openai.com/en/articles/20001339-creating-and-managing-chatgpt-sites) ## The page that is not a toy Private by default, on your own domain, opened by a client with no account, published by any agent over MCP. [ Sign up free → ](https://app.stacktr.ee/?join=1&from=cta_blog_chatgpt_sites_examples) install npx stacktree-install Copy Private hosting for the HTML your agents make. [](https://betalist.com/startups/stacktree?utm_campaign=badge-stacktree&utm_medium=badge&utm_source=badge-featured)[Featured on](https://devhunt.org/tool/stacktree)[](https://buildlist.io)[Get started free](https://app.stacktr.ee/?join=1) ## Product - [Example deliverables](/examples) - [Custom domains](/custom-domains) - [Client feedback](/client-feedback-loop) - [One-time links](/one-time-view-links) - [See how pages get read](/page-engagement) - [Made with Stacktree](/made-with) - [Security](/security) - [Watch the demo](/demo) ## Agents - [All integrations](/agents) - [Claude Code](/claude-code) - [OpenAI Codex](/codex) - [Cursor](/cursor) - [Claude.ai connector](/claude-ai-connector) - [MCP server](/mcp-publish-html) - [Deploy from Claude Code](/deploy-html-from-claude-code) - [Skills](/skills) - [Slack app](/slack) - [n8n node](/n8n) - [Agent payments (x402)](/x402) ## Alternatives - [All comparisons](/alternatives) - [Head-to-head comparisons](/compare) - [Tiiny Host](/tiiny-host-alternative) - [GitHub Pages (private)](/github-pages-private-alternative) - [Vercel](/vercel-alternative-for-agents) - [ngrok](/ngrok-alternative-for-html) - [Display.dev](/display-dev-alternative) - [Static.app](/static-app-alternative) - [OpenAI Codex Sites](/openai-codex-sites-alternative) - [here.now](/here-now-alternative) - [Shippage](/shippage-ai-alternative) - [Best private hosting](/best-private-html-hosting) ## Use cases - [All use cases](/use-cases) - [Share with clients](/share-with-clients) - [Send a file to a client](/share-html-file-with-client) - [Share Claude artifacts](/share-claude-artifacts) - [Share Jupyter notebooks](/share-jupyter-notebook-html) - [Host Storybook privately](/host-storybook-privately) - [Architecture diagrams](/share-architecture-diagrams) - [AI-generated reports](/host-ai-reports) - [Internal HTML tools](/internal-tool-hosting) - [Private HTML hosting](/private-html-hosting) - [Vibe-coded page hosting](/vibe-coding-hosting) - [Leave a website builder](/website-builder-migration) ## Learn - [Blog](/blog) - [Glossary](/glossary) - [FAQ](/faq) - [Agent-loop hosting](/agent-loop-hosting) - [Why agents need a publish primitive](/blog/why-agents-need-a-publish-primitive) - [MCP servers explained](/blog/mcp-servers-explained-for-developers) - [What changed in the 2026-07 MCP spec](/blog/mcp-2026-spec-changes) - [Sites in Codex explained](/blog/sites-in-codex-explained) - [Private-by-default hosting](/blog/private-by-default-html-hosting) - [An agent paid us $1 (x402)](/blog/agent-paid-to-provision-itself) - [When a loop hits a paywall](/blog/loop-engineering-paywall) - [Pricing](/pricing) - [Self-host (new)](/self-host) - [Changelog](/changelog) - [Docs](https://stacktr.ee/docs) - [About](/about) © 2026 Stacktree · stacktr.ee [Privacy](/privacy)[Terms](/terms)[Security](/security)[Dashboard](https://app.stacktr.ee)[npm](https://www.npmjs.com/package/stacktree-mcp)[Sitemap](/sitemap.xml)[llms.txt](/llms.txt)[API spec](/openapi.json) --- Full markdown summary of the Stacktree marketing surface: https://stacktr.ee/llms-full.txt