0x Loom puts an editor, AI helper, GitHub and notebooks in one browser tab

What 0x Loom is

As dev.to reported, 0x Loom is a coding workspace that lives entirely in a browser tab. It combines a code editor, an AI assistant that works alongside you, GitHub tools, notebooks and a store for add-ons. The builder's pitch is that you open the page, drag a folder onto it and start editing.

The editor offers syntax highlighting (colour-coding that makes code easier to read), auto-closing tags and autocomplete. Your files stay in your browser. If you reload the page, your tabs, unsaved edits and open folders come back.

The post describes this as a first public release. Its stated aim is to shorten the path from an idea to running code, so it works on a laptop, a borrowed machine or a phone.

An assistant that asks before it edits

The built-in assistant, which the post calls a Gemini assistant, reads the file you have open and any text you have selected. You can also point it at other files by typing @ followed by a file path.

When it wants to change your code, it does not dump a block of new code for you to paste. It proposes a precise search-and-replace edit, shown as a diff, which is a side-by-side view of what would change. You then press Accept or Reject.

There is a safeguard too. If the text the assistant wants to replace is missing or matches more than one place, the edit is refused. That means a vague or mistaken suggestion should not quietly alter your work.

The assistant can sit behind several API keys, which are the access codes used to call an AI service. When one key hits its usage limit, Loom switches to the next. An AI Billing screen shows requests and tokens, the units AI services count usage in, for each key, so you can see a quota running low.

From first file to live website

You connect GitHub once. After that you can clone an existing repository or start a new one from a Node.js, React (Vite) or Angular template. Loom creates the starting files and pushes the first commit, which is a saved snapshot of the project.

Before each push, changes are listed file by file. Build output and the node_modules folder, which holds downloaded packages, are never sent.

For sharing, a single button publishes to GitHub Pages, GitHub's free site hosting. If the project has a build script, Loom adds the deployment workflow, watches the build and gives you the live address.

Python and Node code runs on a backend. HTML and SVG previews appear next to the code, and a terminal handles npm, the tool that installs JavaScript packages.

Notebooks, a vault and extensions

A Jupyter-style notebook mixes Python and Markdown cells. Jupyter is a popular format for running code in small steps alongside notes. Loom reads and writes standard .ipynb files and shares variables between cells. You can save to your browser, to disk or to a private GitHub repository.

A vault called Tijori stores secrets such as keys. You can drop one into a file at the cursor, which keeps keys out of your source code.

Extensions are plain JavaScript and use a documented interface for commands, panels, status items, key handlers and the AI. An Extension Studio includes a code editor, a live preview of the README, version history and a safety scan. The scan flags vault, network and file access before anyone installs an extension. Publishing is tied to a key only the author holds, and others can fork a copy instead of overwriting the original.

• 3 project templates: Node.js, React (Vite) and Angular
• 15 extensions pre-installed on first run
• 11 colour themes in the marketplace
• 600 files loaded when you clone a repository

What to keep in mind

This is a launch post written by the builder, so the claims are untested by anyone else. The post does not give a price. It also does not say how long files, notebooks or secrets are kept beyond what is stored in your browser.

Python and Node run on a backend, so it is fair to ask where your code goes when it runs. Until that is clear, keep private company code out of it and try it on a throwaway project.

The post ends by asking for feedback on what broke, what is missing and which extension people would build first. That suggests an early-stage product.