Timeline for ChatGPT and Gemini: folders, prompts and a conversation timeline as a browser extension
ChatGPT & Gemini：Add Folders, Timeline & Prompts. - 为 ChatGPT & Gemini 添加文件夹、提示词库和时间轴。
At a glance
- What is it?
- Timeline is a Manifest V3 browser extension that adds a question timeline, saved folders, a prompt library and export to the chat pages of ChatGPT, Gemini, DeepSeek and nine other AI platforms. It is aimed at people who live inside long conversations and keep losing things in them; the trade-off is that everything depends on those pages keeping the DOM shapes the extension reads.
- Who is it for?
- Adopt Timeline if your work happens in long ChatGPT, Gemini or DeepSeek threads and you want navigation, saved folders and a prompt library without moving your conversations anywhere. Skip it if you need a stable programmatic API over your chat history, or if you are not comfortable with GPL-3.0 terms on a codebase you would fork.
- Can I use it commercially?
- Check first. The repository uses a licence we do not classify automatically, so read its LICENSE file before any commercial use.
- Is it still maintained?
- Yes. The repository last received commits 5 days ago.
- What is it written in?
- Mainly JavaScript, according to GitHub's language statistics.
Answers come from the project's GitHub data, last synced on September 15, 2026, and from our analysis. They are not legal advice.
DEEP OPEN-SOURCE ANALYSIS
What Timeline adds to a ChatGPT or Gemini conversation
The README frames the problem in terms of scrolling: to revisit an earlier question you scroll up until you find it, useful conclusions end up scattered across sessions, and formulas or code are awkward to lift out of the page. Timeline answers that by injecting a right-hand timeline into the conversation view, where each node corresponds to one user question. Clicking a node jumps to it, and the README also lists keyboard navigation, a back-to-bottom control and question timestamp labels.
The audience is narrow and specific. This is for people who run long threads on ChatGPT, Gemini, DeepSeek, Claude, Kimi, Doubao, Tongyi Qianwen, Qwen international, Grok, Perplexity, Tencent Yuanbao, ERNIE Bot or NotebookLM, and who want a filing layer on top of them. It is not a client, not a proxy and not an API wrapper. The extension does not replace the platform interface; the README says it only adds the capabilities long conversations lack.
How the extension is put together
The repository is a plain extension directory: manifest.json at the root, with _locales, assets, icons, images, js, popup, scripts and styles alongside it. There is no bundler. The README states directly that this is a browser extension project with no front-end build step, which is why development means loading the source directory rather than producing a dist folder.
The manifest declares Manifest V3 and four permissions, each with a stated purpose. storage holds favourites, folders, prompts, settings and local data. unlimitedStorage exists so that long-conversation favourites, notes and backups are not constrained by the browser's default quota, which is a real concern once you are storing many saved conversations as structured records. identity is used only for Google Drive authorisation. activeTab enables the extension on the AI page currently open in front of the user.
Data flow is local-first. The README says core data lives in the browser by default: favourites, folders, prompts, settings, time labels and notes. Google Drive sync is optional and only runs after the user authorises it, for backup and restore. The project states it does not collect, upload or share conversation content or personal information, and points to the repository as the place to audit that.
One piece of the codebase is imported rather than original. The Gemini image watermark removal engine under js/watermark/engine/ is ported from GargantuaX/gemini-watermark-remover under MIT, and its reverse alpha blending algorithm and calibration masks come from allenk/GeminiWatermarkTool, also MIT. That directory carries its own LICENSE file. The rest of the project is GPL-3.0, and the README notes it continues development of chatgpt-conversation-timeline by Reborn14.
Installing Timeline and saving your first conversation
The README recommends installing from the official browser stores rather than from source. Chrome, Microsoft Edge and Firefox each have a listing, and the README points DeepSeek Harness users to a separate native plugin, dsh-timeline, in its own repository. After installation you open a supported AI conversation page and the features are active; the README says most are on by default and can be turned off per platform or per feature in the extension settings panel.
If you want to run the repository itself, the README gives the unpacked-extension route for Chrome or Edge. Open the extensions management page, enable developer mode, choose Load unpacked, and select the repository root.
# No build step. Load the repository root as an unpacked extension:
# 1. Open chrome://extensions or edge://extensions
# 2. Enable Developer mode
# 3. Click "Load unpacked" and select this repository's root directoryAfter editing code, reload the extension from the management page and refresh the AI platform tab. For Firefox the README names about:debugging as the temporary loading path.
The first useful action is to open a long conversation and check the right-hand timeline. Each node should correspond to one of your questions, and clicking a node should scroll the page to it. From there, save a single question and answer or a whole page into a folder, then open the prompt library and confirm your saved prompts are stored. The README lists JSON import and export plus Google Drive backup and restore under data management, which is the path for moving the same data to another browser or machine.
Where Timeline stops being the right tool
The platform support table is the most informative page in the README, because it is a grid of gaps. ChatGPT, Gemini and DeepSeek get the full set of columns. Claude has no input animation and no question timestamps. Grok has no timestamps, no sidebar favourites and no back-to-bottom. Perplexity and Tencent Yuanbao have no timestamps or sidebar favourites. ERNIE Bot has no smart input and no input animation. NotebookLM only gets highlighting and smart input.
That table is a direct consequence of the design. The extension reads and augments pages it does not control. When a platform changes its markup, the features tied to that markup degrade, and the README does not document a compatibility layer, a versioned selector strategy or a rollback path for a broken release. If a feature you depend on sits in a column marked with a dash for your platform, there is nothing to configure; the capability is simply absent there.
The second boundary is durability. Data lives in browser storage, and the multi-device story is an export file or an authorised Google Drive backup. There is no server-side account, no API and no queryable archive. If you want to search your entire history programmatically or feed it into another tool, an extension that renders a sidebar is the wrong layer. Likewise, if you need a guaranteed record of a conversation for compliance reasons, a browser-local store that the user can clear is not that record.
There is also a licence boundary. The project is GPL-3.0, and the README does not describe a separate commercial licence or an exception. For a team that wants to fork and ship a modified closed-source version, that is a constraint to resolve with your own legal review before writing code.
Timeline versus a prompt manager or a chat export tool
The obvious comparison is with the tools people already reach for: a standalone prompt manager, or a script that exports conversations to Markdown or JSON. Those solve one slice each. A prompt manager gives you a searchable library but knows nothing about the page you are reading; an export script gives you a file but not a way to jump around a live thread.
Timeline's approach is to sit inside the page and treat the conversation as the primary object. That is why the timeline nodes map to user questions rather than to messages, why the question list exists as a separate view for reviewing a thread by question, and why folders, notes and search are attached to saved question-and-answer pairs or whole pages rather than to standalone prompt records. The prompt library is one panel inside that same frame, not the product.
The cost of that approach is coupling. A prompt manager keeps working when ChatGPT ships a redesign. Timeline's timeline, question list and sidebar favourites are the parts most exposed to a redesign, and the platform table shows the project already makes per-platform compromises rather than promising uniform behaviour. If your main need is a prompt library that survives interface churn, a standalone manager is the lower-risk choice; if your main need is to move around and file long threads as you read them, Timeline is doing something the other category does not attempt.
Maintenance, releases and the GPL-3.0 question
The repository is not archived and the last push was on 2026-09-10, four days before this writing, so the codebase is being touched. The README does not publish a release history, a versioning scheme or a deprecation policy, and no releases were retrieved, so there is no changelog to read before upgrading. In practice that means upgrades come through the browser store listings, and the unpacked route is the only way to pin a specific commit. If you run from source, git is your version control and your rollback mechanism; the README does not document one built into the extension.
The upgrade cost is concentrated in the selector layer. Because the extension augments third-party pages, a platform redesign can break features without any change in this repository, and the fix has to land here. Budget for that if you depend on it daily.
On licensing: the project is GPL-3.0, and the README links to the LICENSE file at the repository root. The watermark engine under js/watermark/engine/ is separate, ported from MIT-licensed sources, and carries its own LICENSE file in that directory. The README does not state how the two interact for redistribution purposes. If you plan to ship a modified build, read both files and get your own advice; nothing here should be read as legal guidance.
Editorial conclusion
Adopt Timeline if your work happens in long ChatGPT, Gemini or DeepSeek threads and you want navigation, saved folders and a prompt library without moving your conversations anywhere. Skip it if you need a stable programmatic API over your chat history, or if you are not comfortable with GPL-3.0 terms on a codebase you would fork. Before relying on it, load the unpacked repository in Chrome, open one long conversation on each platform you actually use, and confirm the timeline nodes, the folder sidebar and a JSON export all behave on those pages; the platform support table in the README is the honest scope statement, and the per-platform gaps in it are where you will notice missing features first.
Frequently asked questions
Which AI platforms does Timeline support?
The README lists ChatGPT, Gemini, DeepSeek, Claude, Kimi, Doubao, Tongyi Qianwen, Qwen international, Grok, Perplexity, Tencent Yuanbao, ERNIE Bot and NotebookLM. Support is not uniform: the platform table shows that features such as question timestamps, sidebar favourites, smart input and input animation are missing on some of them.
Where does Timeline store my conversations and favourites?
The README states that core data is kept in the browser by default, including favourites, folders, prompts, settings, time labels and notes. Google Drive sync is optional and only runs after you authorise it, for backup and restore.
Does Timeline send my chat data to a server?
According to the README, the extension does not collect, upload or share conversation content or personal information. The only external path described is the optional Google Drive authorisation, which the README says is used for backup and restore of extension data.
How do I install Timeline from the repository instead of the store?
The README says there is no front-end build step. Open the Chrome or Edge extensions page, enable developer mode, choose Load unpacked and select the repository root; for Firefox it names about:debugging for temporary loading.
What licence is Timeline released under?
The README states the project is GPL-3.0 and links to the LICENSE file at the repository root. The Gemini watermark engine under js/watermark/engine/ is ported from MIT-licensed projects and has its own LICENSE file in that directory.
Can I use Timeline on DeepSeek Harness?
The README says DeepSeek Harness is supported through a separate native plugin called dsh-timeline, hosted in its own repository, which provides the timeline, folders, prompt library, conversation export and formula copying.
Community notes