sandbox:/ mnt/ data/ index.html Not Opening? How to Open the Page ChatGPT Made
Tapping sandbox:/
Short answer
sandbox:/mnt/data/index.html is the web page ChatGPT wrote inside its temporary code environment on OpenAI's servers. It is not on your computer and it is not a web address, so typing it into a browser does nothing. Click the link in the chat while it still works to download the file, then double-click it to open it. If the link has expired, ask ChatGPT to print the whole page as one code block and paste that into a new site to get a live URL.
Turn that index.html into a live link
Paste the code ChatGPT printed, or drop in the downloaded index.html, and get a live HTTPS address in about 30 seconds, with your own domain one step away. Works with every ChatGPT plan. 14 days free, no card.
You asked ChatGPT for a web page and it answered with something like "Your page is ready: sandbox:/". You click it and nothing happens, or you get File not found. You paste it into your browser's address bar and the browser does not know what to do with it. The page exists, just not anywhere you can reach directly. Here is what the path means, how to open the page on your own computer, and how to turn it into a link you can send to someone.
What is sandbox:/mnt/data/index.html?
When ChatGPT writes and runs code for you, it does that in a temporary Python environment on OpenAI's servers: a small Linux machine with its own filesystem. / is the working folder of that machine, and index.html is the file ChatGPT saved there. index.html is the conventional name for the main page of a website, which is why ChatGPT picks it for a landing page, a portfolio or a small tool.
sandbox: is how ChatGPT refers to that machine in its answer. The ChatGPT interface turns sandbox:/ into a download link for as long as the file still exists. It is not a web address and not a folder on your computer. That is why:
- typing it into your browser does nothing:
sandbox:is not a scheme a browser knows; - searching your disk finds nothing: the file only reaches your computer when you download it;
- everybody sees the same path: every ChatGPT code environment uses the same folder layout.
Other pages work the same way. sandbox:/, report.html or landing.html are all files in that same folder, and / without the prefix is the same file as it appears in the code ChatGPT ran. For links to ZIPs, PDFs and other files, and every error message they give, see sandbox:/mnt/data link not working.
How to open the page on your computer
1. Click the link while it still works. In the chat, the link downloads index.html to your Downloads folder. Do it straight away: the code environment is temporary and the link stops working after a break (more on that below).
2. Double-click the downloaded file. An HTML file opens in your browser like any web page. The address bar then starts with file://, which means the page is coming from your own disk, not from the internet.
3. If it looks broken, ask for one self-contained file. A page that shows plain text without colours, or empty boxes where pictures should be, refers to other files (styles.css, script.js, images) that stayed behind in the sandbox. Ask ChatGPT:
Give me the page again as a single index.html, with all CSS and JavaScript inside the file and no other files needed. Print it as one code block.A single file is also the easiest thing to put online later.
On your phone
On a phone the download step is the weak spot. A downloaded index.html often opens in a file viewer that shows the code as text, or it is saved somewhere you cannot easily find again. The route that works on every phone skips the file:
- In the same chat, send: "Print the complete index.html as one code block, with all CSS and JavaScript inside the file. Only the code, no explanation."
- Tap Copy on the code block in the answer.
- Paste it into a new site (see From index.html to a live link). You get an
https://address that opens in your phone's browser like any website, and that you can send to someone else.
sandbox:/index.html, dashboard.html and other names
ChatGPT does not always write the full path. sandbox:/, sandbox:/, sandbox:/, quiz.html, report.html or landing.html are all the same kind of thing: one HTML file in the temporary environment of your chat. Everything on this page applies to each of them. Only the file name in your request changes, for example "print the complete dashboard.html as one code block". The short form has its own page: sandbox:/index.html does not open.
When the link says File not found
The code environment is temporary. After a while without use it is reset, and everything in / disappears with it. OpenAI does not publish how long ChatGPT keeps it; for the code interpreter in its API, OpenAI documents that a container expires after 20 minutes without use and that its data is then discarded. Clicking an old link gives "File not found" or "The code interpreter tool session has expired." Sometimes the link never worked at all, because the code that should have written the file did not run in that turn.
The page is not lost as long as the conversation still holds its code. Two ways back:
- Ask for the code as text. Ask ChatGPT to print the complete
index.htmlin one code block. This does not depend on the sandbox, so it always works. Copy the block, paste it into a text editor and save it asindex.html, or skip the editor and paste it straight into a new site (next section). - Ask for a fresh link. Ask ChatGPT to recreate
index.htmlin/, list the folder so you can see the file exists, and give you a new link. Download it immediately.mnt/ data
Already downloaded the page and stuck with an address like file:///? That link only works on your own computer: see why nobody else can open it.
Not sure the code is the right page? Paste it into the free HTML preview first: it shows the page in your browser, without an account.
Turn that index.html into a live link
Paste the code ChatGPT printed, or drop in the downloaded index.html, and get a live HTTPS address in about 30 seconds, with your own domain one step away. Works with every ChatGPT plan. 14 days free, no card.
From index.html to a live link
A downloaded index.html only works on your own computer: the file:// address means nothing to anyone else. To send the page to someone, it has to be on a web host. Two routes:
ChatGPT Sites. On Plus, Pro and workspace plans, ChatGPT can publish a site on OpenAI's hosting. OpenAI's help center says Sites is not available on Free or Go.
Your own host, from any ChatGPT plan. On VibeDeploy:
- Start the 14-day trial (no card) and create a site.
- Choose Paste code and paste the code block ChatGPT printed, or drop in the downloaded
index.html. About 30 seconds later the page is live athttps://.name.subsite.site - Add your own domain in the dashboard and create the DNS records it lists. The certificate is issued automatically.
Paste the code itself, which starts with <!DOCTYPE html> or a tag such as <div>, not the sandbox:/ line: that line only names the file, it does not contain it.
If ChatGPT made several files, ask for one ZIP and drop that in instead. If the ZIP holds a package.json (a React, Vite or Next.js project), it is source code rather than a finished page: ask ChatGPT for a plain HTML version, or build it yourself with npm run build and upload the build folder, usually dist/. More routes, plan by plan, in deploy a ChatGPT-built website.
Using Grok? Its sandbox path is /home/workdir/artifacts/index.html. Claude's is /mnt/user-data/outputs. The list of AI sandbox paths has the rest.
Sources
- OpenAI API documentation, Code Interpreter: "A container expires if it is not used for 20 minutes", after which its data is discarded and not recoverable. This covers the API; OpenAI publishes no timeout for ChatGPT itself. Checked 26 October 2026.
- OpenAI Developer Community, Issue with PDF download link in ChatGPT, "sandbox:/mnt/data" not accessible (opened 3 April 2025): sandbox links that do not download or say "Not found", and regenerating the file for a fresh link as the fix that worked. Checked 26 October 2026.
- OpenAI Developer Community, Download link not working for generated files in ChatGPT code interpreter (opened 5 April 2025): "file not found" on generated files. Checked 26 October 2026.
- OpenAI Help Center, Creating and using ChatGPT Sites: Sites on Plus, Pro and workspace plans, not on Free or Go. Checked 13 October 2026.
Ship your AI-built site in minutes
VibeDeploy hosts your AI-built websites in the EU with custom domains, automatic SSL, and a 14-day free trial that gets you online today.
Deploy guides for this topic
Questions people ask
What is sandbox:/mnt/data/index.html?
It is a web page, index.html, that ChatGPT created in the working folder /mnt/data of its temporary code environment, a small Linux machine on OpenAI's servers. The sandbox: prefix tells the ChatGPT interface to turn the link into a download while the file still exists. The file is not on your computer until you download it.
Why does nothing happen when I open sandbox:/mnt/data/index.html in my browser?
Because sandbox: is not a web address. Only the ChatGPT interface understands it, and only for as long as the code environment that holds the file is alive. Your browser and your own disk have no such location, so pasting the path into the address bar or searching for it finds nothing.
How do I view the index.html ChatGPT made?
Click the link in the chat to download the file, then double-click the downloaded index.html: it opens in your browser like any web page. If the page looks unstyled or images are missing, it refers to other files (a CSS file, a script, pictures) that you did not download. Ask ChatGPT for a single index.html with the CSS and JavaScript inside it.
The link says File not found. Is my page gone?
The copy in the sandbox is, but the page is not lost as long as the chat still has the code. Ask ChatGPT to print the complete index.html as one code block. Copy it and save it as index.html, or paste it straight into a new VibeDeploy site. Asking it to recreate the file and give a new link also works, if you download it right away.
How do I share the page with someone or put it online?
A downloaded index.html only works on your own computer. To share it, upload it to a web host. On VibeDeploy you paste the code or drop in the file and get a live HTTPS address in about 30 seconds, then add your own domain. On Plus, Pro and workspace plans, ChatGPT Sites can also publish a site on OpenAI's hosting.
How do I open sandbox:/mnt/data/index.html on my phone?
Skip the download: on a phone a saved .html file often opens in a file viewer as text instead of as a page. Ask ChatGPT to print the complete index.html as one code block, copy it, and paste it into a new VibeDeploy site. You get an https link that opens in any phone browser and that you can send to others.
Is sandbox:/index.html the same thing?
Yes. ChatGPT sometimes shortens the path or picks another file name: sandbox:/index.html, sandbox:/mnt/data/dashboard.html, game.html, quiz.html or report.html are all files in the same temporary environment. The steps are the same for each of them.
Related reading
sandbox:/ index.html Does Not Open? How to See the Page ChatGPT Made
ChatGPT answered with sandbox:/
AI Sandbox File Paths: Where ChatGPT, Claude, Grok and Others Save Your File
/
/ home/ ubuntu/ index.html: Where Manus Saved Your Site
Manus says your site is at /
/ mnt/ agents/ output/ : Where Kimi Saved Your File
Kimi says your page is at /