Self-host Excalidraw
- License
- MIT
- GitHub stars
- 135,303, as of 8 Nov 2026
- Latest release
- v0.18.1, as of 8 Nov 2026
- Host needs
- Not stated in official docs
- Install method
- Docker
- One-click template
- Available
- License: MIT.
- Last commit: 6 Nov 2026 01:10 SGT, as of 8 Nov 2026. The repository is not archived.
- GitHub stars, as of 8 Nov 2026: 135,303. Latest release, as of 8 Nov 2026: v0.18.1, published 21 Apr 2026 16:43 SGT.
- Self-host limit: the docs say "At the moment, self-hosting your own instance doesn't support sharing or collaboration features."
Requirements
CPU, RAM and disk are not stated in the official docs we read.
The one stated dependency is Docker: the self-hosting section gives docker build and docker run commands for the client image.
Docker or Docker Compose
The official guide is the "Self-hosting" section of the Excalidraw development page.
It says the project publishes "a Docker image with the Excalidraw client at excalidraw/excalidraw" and that you can run it "under your own domain, on Kubernetes, AWS ECS, etc."
Hosting options
A Railway template for Excalidraw is at railway.com/deploy/excalidraw, created by Tasha.
It was created on 22 Jun 2026, builds from the repo Amritasha / excalidraw-railway, and showed "Total Projects 4" on 17 Oct 2026.
The docs give no RAM figure, so the template cannot be checked against the Railway Free limit of "0.5 GB RAM per service".
Affiliate disclosure: Forkshelf earns a commission if you sign up for Railway through the links marked "Paid link" on this page. This does not change the order or content of any guide here.
To get started, create a Railway account (Paid link).
Backups and updates
The self-host doc covers only the client image.
A backup method is not documented on the pages we read, and neither is an update method.
Alternatives
Excalidraw is listed on open-source alternatives to Figma and Miro.