What I Built Instead
I recently read an article about a professor who lost years of work on Medium. Overnight, content he assumed was permanent became inaccessible. No warning. No clean recovery path. Just gone.
That story hit close to home—because a quieter version of the same thing has already happened to many of us using AI platforms. Our conversations, research, drafts, and long-form thinking live inside someone else’s system. Even when exports are available, they’re often dumped into massive JSON files that are technically “yours” but practically unreadable. Ownership without usability isn’t ownership at all.
I realized something uncomfortable: my AI conversations are work product, not disposable chat logs. They include research, writing, legal notes, code, and ideas I may want years from now. Relying on a platform’s goodwill or UI to access that history felt no different than trusting any other centralized publishing platform.
So I did what I usually do when a tool doesn’t exist.
I built one.
The Problem With AI Exports (Nobody Talks About This Part)
Most AI platforms let you export your data. That sounds reassuring—until you actually try to use it.
What you usually get is:
- a massive
conversations.jsonfile, - maybe a static HTML file,
- no search,
- no browsing,
- no way to read individual conversations cleanly,
- no way to host it yourself in a usable format.
In other words: you technically own the data, but you can’t really use it.
That gap is where people lose work—not because the data vanished, but because it became inaccessible.
The Solution: A Self-Hosted ChatGPT Archive Viewer
I built a lightweight, self-hosted viewer that lets you:
- upload your ChatGPT export (
conversations.jsonor ZIP), - browse conversations in a clean sidebar,
- click and read full threads instantly,
- download individual conversations as clean HTML,
- print or save them as PDF,
- host everything on your own server or WordPress site.
No external APIs.
No tracking.
No subscriptions.
No platform lock-in.
Just your data, readable again.
How It Works (High Level)
- The viewer runs as a small PHP + JS app.
- It reads your exported
conversations.json. - Conversations are parsed and displayed like an email or chat client.
- Each conversation can be exported individually as HTML (PDF via print).
- You control where it lives: standalone, WordPress, or embedded in Framer.
This isn’t a replacement for ChatGPT.
It’s an insurance policy for your thinking.
WordPress + Framer Embed Snippet
WordPress (shortcode)
After installing the plugin, drop this into any page:
That’s it. The viewer renders inside the page automatically.
Framer (Embed component)
Host the viewer on your domain, then embed it in Framer with:
<iframe
src="https://yourdomain.com/chatgpt-archives/"
style="
width:100%;
height:85vh;
min-height:700px;
border:0;
border-radius:14px;
overflow:hidden;
"
loading="lazy">
</iframe>
This keeps all processing on your server while Framer handles layout and presentation.
One-Paragraph Install README (Simple & Honest)
Installation:
Unzip the plugin folder and upload it to wp-content/plugins/. Activate “ChatGPT Archives Viewer” from the WordPress admin panel. Create a page and insert the shortcode
. To import data, upload your exported conversations.json (recommended) or the full export ZIP through the viewer’s upload screen. For security, uploads should be restricted to site admins. The viewer runs entirely on your server and does not send data to third parties.
Why This Matters (The Bigger Point)
Platforms change.
Policies change.
Accounts get flagged.
Products get shut down.
Your thinking shouldn’t disappear just because a UI does.
AI conversations are becoming research notes, drafts, journals, and intellectual property. Treating them like disposable chat history is a mistake we’ll all regret later.
This tool exists for one reason:
to make sure your work survives the platform it was created on.
If you export your data, you should be able to read it.