Quick start

Launch latest Qiskit (…) in your browser Opens latest-xl on mybinder.org. First start is slow (image pull); subsequent launches are cached.

Sessions are temporary. Nothing is saved: download your notebooks (File → Download) before you close the tab. An idle session shuts down after about 10 minutes.
Memory is limited (1–2 GB), so large simulations can restart the kernel.
Stuck on “Waiting” for more than 3 minutes? Reload the page; mybinder tries another server. mybinder’s user guidelines

New here? Open a ready-made notebook in your browser:

Qubits, gates and circuits in Qiskit From IBM’s Basics of Quantum Information course. Runs entirely in your browser session; no account needed. Quantum Coin Game An interactive slideshow from fun-with-quantum.org. Run on a real quantum computer Qiskit’s intro tutorial. Needs a free IBM Quantum account for the hardware step; on mybinder, paste your key inline instead of save_account() (see below).

Or run locally with Docker:

docker run --rm -p 8888:8888 -v "$PWD:/home/jovyan/work" ghcr.io/qubins/images:latest-small

Open the http://127.0.0.1:8888/lab?token=… URL it prints. Save your notebooks in work/: that is the folder you ran the command in. Everything else is deleted when the container stops.

Podman, Windows, a busy port

Podman: the same command with podman instead of docker.
Windows PowerShell: -v "${PWD}:/home/jovyan/work"; cmd.exe: -v "%cd%:/home/jovyan/work".
Port 8888 already in use: -p 8899:8888, then open port 8899 instead.

Launch your repo or notebook on QuBins

Which flavor? xl for tutorials, addons + scientific stack (the usual pick). small for a lean image with just core Qiskit (bare launch or Docker; it can’t load a repo or notebook URL, so this form offers xl images). xxl adds the local AI transpiler (PyTorch + CUDA, ~4 GB, amd64); for now only 2.4-xxl has it, because the transpiler does not support Qiskit 2.5 yet.

Which tag? latest-xl and the newest minor (now 2.5-xl) are kept warm on mybinder, so readers start fastest. Other tags launch too, just slower the first time on each server. For a frozen environment (a course, a release), use a monthly snapshot such as 2.5-xl-20261008; it is never changed and kept for 12 months. More

Paste a GitHub repo URL (or a /blob/ URL, or a raw .ipynb URL). You get a launch link to share, and an optional Markdown badge for your README.

Launch link (share this; it goes through qubins.org/launch/ to mybinder.org)

Badge for your README (optional — paste into a Markdown README; click the badge to test)

Why does this go through /launch/?

The link and the badge point at our /launch/ redirector rather than straight to mybinder.org. When mybinder's URL form changes (it has, historically), every link already in the wild keeps working with no edit needed on the reader's side. It also catches a mistyped or retired image tag and offers a working one.

All images

Every image is daily-rebuilt, Trivy-scanned (HIGH/CRITICAL gated), and cosign-signed via GitHub Actions OIDC. The Docker tag template for every row is: docker run --rm -p 8888:8888 -v "$PWD:/home/jovyan/work" ghcr.io/qubins/images:<tag>

small vs xl vs xxl — full comparison

	small	xl	xxl
Use for	Lean image, fast pull, core Qiskit work	Tutorials, docs notebooks, addons, scientific stack (the usual pick)	Everything in xl plus the local AI transpiler — currently `2.4-xxl` only (`2.5-xxl` has the same content as `2.5-xl` until the transpiler supports Qiskit 2.5)
Download	~0.4 GB	~0.9 GB	~4 GB (amd64)
Included	`qiskit` `qiskit-aer` `qiskit-ibm-runtime`	Qiskit ecosystem: `qiskit[all]`, all `qiskit-addon-`, `qiskit-experiments`, `qiskit-serverless`, `qiskit-ibm-catalog` Scientific stack:* scipy, sklearn, pyscf, plotly, sympy, ffsim, pandas Notebook tooling: `pylatexenc`, `nbgitpuller`, `jupyterlab-open-url-parameter`	Everything in xl, plus `qiskit-ibm-transpiler[ai-local-mode]` (pulls PyTorch + the full CUDA 13 wheelset)
Load a repo or notebook URL	— (bare launch only)	✓ (via `jupyterlab-open-url-parameter`)	✓ (inherits from xl)
arm64 caveats	none	none from 2.5 on (packages without aarch64 wheels are built from source); 2.0–2.4 omit `gem-suite`	the AI transpiler is amd64-only (no aarch64 wheels), so the arm64 build equals xl

For slideshows with interactive widgets, 2.5-xl also ships the classic Notebook frontend with classic RISE: launch it with ui=rise-classic, as the Quantum Coin Game above does. The older 2.1-xl-rise is kept for existing links.

Filter: Qiskit version Flavor

Qiskit	Flavor	Launch	Docker tag

How it works

Dockerfile is parameterised by Qiskit version × flavor and installs the matching versions/<target>/requirements.txt. A GitHub Actions workflow builds per-arch (amd64 + arm64), gates on a Trivy scan, publishes the multi-arch manifest to GHCR with a SLSA provenance attestation, signs it with cosign, and force-pushes a tiny stub branch (the image reference plus a welcome notebook) that mybinder.org consumes to skip the dependency-tree rebuild.

A daily cron rebuilds everything against the current base image to absorb upstream CVE fixes. Dependabot watches the docker base, GitHub Action versions, and the latest Qiskit minor's pip pins.

Why trust these images

Scanned: every build is scanned with Trivy, and an image with a fixable HIGH or CRITICAL vulnerability is not published.
Rebuilt daily: security fixes in the base image and packages arrive within a day.
Signed and traceable: each image is signed with cosign (keyless, via GitHub Actions) and carries a SLSA build-provenance attestation, so you can check it was built from this repository (how to verify).
Multi-arch: linux/amd64 and linux/arm64 (Apple Silicon, Raspberry Pi 5, Graviton).

QuBins — the place for your QuBits: prebuilt quantum compartments, pick one, run your Qiskit notebook on (my)binder or as a container (“bin”).