(If your browser shows an “upload X files” warning, it is only the folder‑picker UI; the script still processes **only** files named conversations.json.)
How to use this viewer
Purpose – This tool is designed to view data exports from ChatGPT offline. You can use it in the web browser (data stored and processed on your local machine), or you can just save this HTML file on your machine and open it locally.
It recursively scans a chosen **root folder**, finds every file named conversations.json (including in sub‑folders), merges the JSON arrays, and renders the conversations in the browser. All rendering happens locally – no network traffic.
Unzip each export into its own folder (e.g. 2024‑01, 2024‑02, …).
Place all those folders inside a single “root” folder.
Open this chatgpt‑viewer.html file, click Select Export Folder, and choose the root folder.
The viewer will gather all conversations.json files and display each chat in a layout very close to the native ChatGPT UI.
Use the **search bar** to find keywords. The search runs entirely offline; it may be slower for very large numbers of conversations or on low‑end hardware.
Limitations
Only textual conversation data is rendered. Image, video, or other multimodal outputs are not displayed.
Requires a modern browser that supports the File System Access API (Chrome, Edge, Opera). Older browsers will fall back to the classic folder picker and may show an upload warning.
Third‑party libraries & licenses
marked – Markdown parser – MIT License. view license
Fuse.js – Fuzzy‑search engine – MIT License. view license
github‑markdown‑css – CSS that mimics GitHub’s markdown rendering – MIT License. view license
Created by Mingfeng Lin.
You may use or redistribute this tool, but you must keep the attribution and license notice. Comments welcome!.