
Colorblind
io.github.vince-gonzalezPublished by GitHub account vince-gonzalezv0.2.0Updated Oct 24, 2026
See a color or image as a colorblind person does; recolor images and make Ishihara plates.
Overview
Local MCP server that lets an assistant simulate colorblind vision, recolor images, check palette safety and generate Ishihara test plates.
- What you can do
- Simulate how a color or image appears under protan, deutan or tritan vision
- Recolor images to show a colorblind view
- Check whether a color palette stays distinguishable, using CIEDE2000 color difference
- Generate Ishihara-style test plates and suggest colorblind-safe palettes
- When to use it
- Useful when designing charts, palettes, maps or interfaces that must stay readable for colorblind users, or when you want to preview an image as a colorblind viewer would see it. Also handy for producing Ishihara-style test plates.
- Requirements
- Runs as a local stdio process, installed from the npm package colorblind-mcp; Node.js is needed. No account, API key or environment variable is declared.
Installation
In SourceWeft
- Open Colorblind in the dashboard and add it to a workspace.
- Enable the server for the chats that should use its tools.
Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.
Other MCP clients
Follow the launch instructions in the repository.
README
An open, published method for color vision — and 17 vision tests built on it. Everything runs on the device. No account, no upload, no image ever sent anywhere.
The colorimetry is not a private formula: it is Machado, Oliveira & Fernandes (2009) deficiency simulation with CIEDE2000 color difference, published open access at doi.org/10.5281/zenodo.21310578, and shipped as packages so you can run the math yourself and check our numbers against yours.
Use the engine
For LLM tools: cvdsafe-mcp and colorblind-mcp are MCP servers, listed in the official
Model Context Protocol registry — so an assistant can check a palette or generate an Ishihara
plate as a native tool.
The tests
Color vision, D-15 arrangement, hue discrimination, anomaloscope-style matching, contrast sensitivity, acuity, astigmatism, Amsler grid, blind spot, vernier alignment, flicker fusion, saturation, dominance, reaction, stereo depth, digital eye strain, and a shape-based test for kids.
A test on a web page cannot be a diagnosis: an uncalibrated display, ambient light and viewing distance all move the result. Each test says so where it reports, and points at an eye examination rather than standing in for one.
Audit it yourself. opticquiz.com/methodology answers the eight questions worth asking of any online vision test — who made it, what it measures, what it assumes about your screen, whether it has been clinically validated (it has not), what it cannot measure, and who pays for it.
Also in here
Translations (de, es, fr, hi, it, pt, zh), a browser extension (reviewed and listed by Mozilla),
a VS Code extension, a Figma plugin, an embeddable widget, and the research notes behind the
plate colors under research/.
MIT licensed. Issues and corrections welcome — if a number here disagrees with yours, that is a bug worth reporting.
Part of F-Keys — independent hardware, software and internet products. See the working log and live status.
Source: README.md at commit 9e2fdfb
Tools
0Version history
1- v0.2.0LatestSep 16, 2026
Score
50% of the applicable checks met
How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.
- Reached in the last 30 days25 pts
Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.
- Auth declared10 pts
Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.
- Tools listed15 pts
Answer tools/list without a credential, or ship a package that starts in a sandbox without one.
- DescriptionRequired15 pts
- Logo10 pts
- Documentation15 pts
- License10 pts