Sharing
Publish a read-only link to a notebook so anyone can view it, fork it, or open it elsewhere.
When you want to put a notebook in front of someone, you can publish a share link: a read-only view that anyone can open in their browser, no account needed. The link is a short clusy.io/s/… URL, and it unfurls with your title, description, and a preview card when you paste it into Slack, X, or LinkedIn. You can also list the notebook on the Clusy Hub for anyone to discover.
Setting up a share
You build a share from the share dialog, which gives you a few choices:
- Scope. Share this experiment (a single branch, as a linear notebook) or the whole Lab (every branch, with the fork explorer so viewers can walk the tree).
- Presentation. Give the share a title and description for the share page and link previews. Generate with AI can draft both from the notebook if you'd rather not write them yourself.
- What viewers can do. Toggle whether viewers can see outputs and results (the charts, tables, and logs frozen into the snapshot), fork it into their own workspace, and deploy it on to Hugging Face, Kaggle, or GitHub.
- List on the Clusy Hub. Off by default. Turn it on to add the notebook to the public Hub gallery, where you can also give it up to eight tags.
The Share & publish dialog: scope, title, description and tags, the viewer permission toggles, and the Hub listing switch
A share is a frozen snapshot
Creating a share freezes the notebook as it is at that moment: the cells, the branch structure, and the saved output. That snapshot is what viewers get.
It does not follow your work. Once the link is out, you can keep editing, run failed experiments, and break things, and none of it reaches the share. Nothing you do after the snapshot leaks.
The flip side is that a share goes stale. When you want viewers to see the current state, reopen the share dialog and hit Update, which re-freezes the notebook into a new snapshot at the same URL. (Saving other changes to the share has an Also re-freeze the current notebook state into the snapshot toggle for the same purpose.) The link never changes; only what it shows does.
Everything in an output ships with the share
The snapshot is a faithful copy of what you saw, which includes whatever was
sitting in a cell's output when you shared it: an API key you printed while
debugging, a token in a stack trace, a customer name in a df.head(), an
absolute path with your username in it. Read your outputs before you share, not
after.
There is no expiry, no password, and no per-person access list. A share link is public or it is nothing: anyone who has the URL can open it. If you need the code out but not the results, turn see outputs and results off. If a link gets out that shouldn't have, revoke it.
What a share shows
A shared notebook shows the cells, the branch structure, and the saved output (tables, charts, all of it), along with the title and description you set. The viewer can read it but can't change it.
A shared notebook in the viewer: the author and date up top, readable cells with their saved output, and actions to download, open in Colab, deploy, or fork it
What a viewer can do
Depending on the permissions you set, a viewer can:
- Fork it to their own workspace and take your notebook as a starting point for their own work (with an account).
- Open it elsewhere by downloading the
.ipynb, or taking it to Colab (the download plus a pointer to Colab, since Colab can't open a URL directly). - Deploy it out to GitHub, Hugging Face, or Kaggle. Deploying works by forking first, so it's available when forking is.
- Build their own, a prompt for people who don't have an account yet.
Downloading the .ipynb is always available on a share, whatever else you toggle off.
Managing share links
You create, update, and revoke share links from the project. Revoking a link turns it off for everyone holding the old URL, and removes the notebook from the Hub if it was listed.
Sharing vs. publishing
A share link is a quick read-only view hosted by Clusy. Publishing is different. It pushes your notebook (and the models or datasets with it) onto Hugging Face or Kaggle under your account there. Listing on the Hub is a third option: the notebook stays on Clusy, but it joins the public gallery.
| You want to… | Use |
|---|---|
| Show someone the result, fast | A share link |
| Let anyone discover your work | The Clusy Hub |
| Version the notebook with a history | GitHub |
| Put a model and notebook on a public hub | Hugging Face or Kaggle |