Gantt charts that a screen reader user can question
gantt-a11y adds keyboard commands and spoken answers to Gantt chart components, so that someone who cannot see the chart can ask what a task depends on, what runs at the same time, and what a change broke.
Status. An open-source extension, checked so far only by its author, a sighted developer: fully with VoiceOver in Safari 26.2 on macOS 26.2, partly with NVDA 2026.3.1 in Chrome 157 on Windows; not with JAWS or Firefox. No feedback yet from people who use a screen reader every day. Nothing here is a claim of conformance with WCAG or any other standard.
The problem
A Gantt chart is a drawing: bars on a time scale and lines between them. A sighted user sees at a glance which task waits for which, what overlaps, and that a move pushed a task past its deadline. Components with keyboard navigation let a screen reader user move from row to row and hear a task's name and dates. The relations between tasks, the part a Gantt chart exists for, stay inside the drawing.
What it does
- Reads the dependencies of the focused task by name, and pages through long lists.
- Moves focus along links to what a task depends on and what depends on it, and back again.
- Answers questions about time: what runs at the same time and for how long, what ends before, what starts after.
- Moves tasks and changes durations from the keyboard, and says the result first, including any link it broke.
- Creates and removes links from the keyboard.
- Says what happened after the task form is saved, cancelled or deleted.
- Gives an overview of the chart, says where the focused task is, and goes from one conflict to the next.
- Speaks its own list of keys, in groups.
The commands live in a small engine that knows nothing about a particular library. Adapters connect it to dhtmlx-gantt (Community Edition) and frappe-gantt, which has no keyboard support of its own.
Who it is for
- Developers who ship a Gantt chart in their application and want it to be usable without sight, using only the library's public API.
- People who plan or follow projects with a screen reader, and the teams they work with.
- Library authors looking for a worked example of what a non-visual Gantt chart needs. The announcement specification and the audit are meant to be reusable.
What it does not do
- It does not make the drawing itself accessible: bars and lines stay pictures. The facts they show are made available as text on demand.
- It does not replace the library's own keyboard navigation; it adds commands on top of it.
- It has no dialogs: everything is spoken through the row label or a live region.
- It has not been checked with JAWS, with Firefox, with keyboard layouts other than English or with braille displays, and NVDA only partly.
- It is not a certification. The automated audit checks keyboard behaviour and what the browser exposes, not what a person can do.
Try it
The two demos have a switch to compare the chart with and without the extension. How to try with a screen reader has the steps for VoiceOver, NVDA and JAWS, and says which of them have been checked.