Accessibility Conformance Report
Measured against WCAG 2.1 Level A and AA, and Section 508.
Last updated: September 13, 2026
Why some of this says “partially supports”
Because it is true. A report claiming full conformance while measurable failures remain is worth less than no report at all, and it is the first thing an evaluator checks. Where Algebro does not yet conform, this says so and gives a date.
This is a self-assessment. The organisation that maintains the reporting template states that these reports need not be third-party reviewed, and this one has not been. What follows is what testing actually found.
What was evaluated, and how
The whole product: the student surfaces where questions are answered, the teacher and school administrator surfaces, the parent portal, and the public pages. Evaluation used two automated methods that see different things, plus a manual review that is scheduled and not yet complete.
- Static analysis of every component against the full accessibility rule set, covering names, roles, labels and keyboard operability.
- Automated testing of the rendered pages while signed in as a student, a teacher, a parent and a school administrator, since most of the product is behind a login and cannot be assessed from the outside. This is what measures contrast, heading structure and landmarks.
- Manual review with assistive technology: scheduled, not yet complete. Automated tools find roughly a third of accessibility problems. Criteria that depend on human judgement are marked “not evaluated” below rather than assumed to pass.
WCAG 2.1 Level A and AA
Perceivable
| Criterion | Conformance | Remarks |
|---|---|---|
| 1.1.1Non-text content | Partially supports | Every mathematical figure carries a generated text description, and questions that ask a student to read a value from a graph deliberately describe the figure without stating that value. Decorative images are hidden from assistive technology. Not yet reviewed with a screen reader by a person. |
| 1.3.1Info and relationships | Supports | Form controls are associated with their labels. Captions that head a set of controls, such as a row of answer choices, are exposed as a named group rather than as a stray label. |
| 1.4.1Use of colour | Supports | Links inside running text carry a persistent underline, not colour alone. Status is given in words as well as colour. |
| 1.4.3Contrast (minimum) | Supports | Automated testing across nine signed-in pages, one per role, reports no text below the 4.5:1 threshold. Pairings were also checked against all fifteen colour themes the product ships, light and dark, rather than only the default. |
| 1.4.4Resize text | Not evaluated | Scheduled with the manual review. |
| 1.4.10Reflow | Not evaluated | The product is responsive to a 390px viewport and is used daily on phones, but conformance at 400% zoom has not been measured. |
Operable
| Criterion | Conformance | Remarks |
|---|---|---|
| 2.1.1Keyboard | Supports | Every interactive control is reachable and operable by keyboard. The floating tools a student gets during a practice test, the calculator, scientific calculator, notepad and reference sheet, can each be moved with the arrow keys, not only dragged with a pointer. |
| 2.1.2No keyboard trap | Not evaluated | Scheduled with the manual review. |
| 2.4.3Focus order | Partially supports | Dialogs trap focus and return it to whatever opened them. Some forms move focus to the first field on open, which is deliberate and expected in context; it is disclosed here rather than removed. Not yet verified by hand for keyboard traps in the drawing tools. |
| 2.4.7Focus visible | Supports | Focus is visibly indicated on every focusable element. |
Understandable
| Criterion | Conformance | Remarks |
|---|---|---|
| 3.1.1Language of page | Supports | The document language is declared. The product is available in English only. |
| 3.3.1Error identification | Supports | Validation errors are given in text, not by colour alone. |
| 3.3.2Labels or instructions | Supports | Inputs carry labels or accessible names. Answer boxes state the form expected, for example whether to enter a fraction. |
Robust
| Criterion | Conformance | Remarks |
|---|---|---|
| 4.1.2Name, role, value | Partially supports | Interactive elements expose an accessible name and role. A small number of composite widgets, notably the graphing tools a student draws on, have not been reviewed with a screen reader. |
| 4.1.3Status messages | Not evaluated | Scheduled with the manual review. |
Things worth knowing that no criterion asks about
Read aloud is built in. A student can have a question read to them, including the mathematics, which is read the way a trained reader reads it on a state test rather than as symbols: “three fourths”, “negative five”, “f of x”, and “the quantity x plus 2” for a group in parentheses. Figures and tables are described too, the way Florida’s own read-aloud describes them: a graph’s axes and scale and what it prints, a table by its headings and then row by row. It is available on every surface where a student answers a question, including the full-length practice test, the diagnostic and the placement test. It speaks with a voice built into the student’s own device, so the question is never sent anywhere to be read. Read-aloud is a standard accommodation on Florida’s state assessment, and this exists so a student entitled to it on the real test can practise with it.
Figures never give the answer away. When a question asks a student to read a value off a graph, the figure’s description says what is drawn without saying where. A student using a screen reader gets the same task as everybody else, not an easier one.
The product is English only. There is no Spanish version and no translation layer.
There are no per-student accommodation settings. Extended time, reduced answer choices and similar supports are not configurable. Where a school needs them, they have to come from the device, the browser, or an adult working alongside the student.
Known issues and when they close
| Issue | Target |
|---|---|
| A small number of composite widgets, chiefly the graphing tools a student draws answers on, have not been reviewed with a screen reader. | December 2026 |
| Manual review with assistive technology, covering reflow, keyboard traps, status messages and the graphing tools. | December 2026 |
| Per-student accommodation settings, such as extended time or reduced answer choices. | Not scheduled |
Tell us what we missed
If something on Algebro is hard or impossible to use, we want the specific page and what went wrong. Reports from people using assistive technology are worth more than any automated test, and they get answered by the person who wrote the code.
jkorwin@algebro.aiSee also our privacy policy.