Frequently asked questions

Why does this exist?

JADE was built by Dr Miriam J Johnson, who wanted to measure the rate of gendered language in publishing industry job descriptions. She developed the tool from an open source foundation and expanded it significantly over time. It is designed to help organisations identify exclusionary language in job descriptions, general text, and manuscripts, and we are always building on these features to improve results for our users.

What does it do?

JADE checks written content for words and phrases that are broadly exclusionary, dividing flags into categories so users can understand where changes are needed and track improvement over time.

How does it work?

JADE uses an open source natural language processing tool to identify hundreds of specific words and phrases in relation to the language around them. The terms and phrases flagged have been identified by Dr Johnson and other academics, and the list is continually reviewed and expanded. The tool can also detect the likely job title from the text itself, in case a user does not fill in that field manually.

Within the job description tool, JADE also flags whether a salary has been included. Salary transparency matters. Tell people what you will pay them.

How you know what the tone is?

Tone analysis looks at patterns in your writing across four dimensions: how formal or casual it is, how welcoming it feels to readers, how easy it is to understand, and whether it treats the reader as an equal. Each dimension is scored out of 10 based on your word choices, sentence structure, and phrasing. The labels accompanying each score describe what that score means in plain English. Tone analysis is processed via Anthropic's API. Anthropic does not use API data for model training, and inputs are not retained beyond processing

How do you know the reading score?

Reading score is determined by analyzing the complexity of the text using a combination of factors including sentence length, word frequency, and syntactic structure. It is based on the Flesch-Kincaid Grade Level and calculator. This helps assess how easy or difficult the text is to read and understand.

How does the report generation work?

Reports are generated using a combination of your submission data and Anthropic's API. For tone analysis, the text you submit is processed by the API to generate your scores. For your action plan, the flagged terms identified in your text are sent to generate plain English guidance. For all other report sections, only aggregated statistics are used: your average score, flag rate, and sector. No submitted text is stored by Anthropic, and Anthropic does not use API data for model training. You can read their privacy policy at anthropic.com/legal/privacy.

What is the difference between no account, a free account, and a paid account?

Full details are on the pricing page. In short: without an account you can run checks and see basic results. With a free account you get five checks per month, and your results accumulate in a personal dashboard so you can see how your texts perform as a group. A paid account gives you everything in the free tier plus expanded dashboard data, charts with adjustable comparison points, downloads of all results and charts, and the ability to generate a full report with summary and action points.

How do you handle my data?

We do not store any text submitted to JADE. Once a check is complete, the content is gone; you cannot re-run a piece without submitting it again from scratch. We store only your email address, name, company name, and organisation size, as you provide them. You can delete your account and all associated data at any time from your account settings. Deletion is permanent and cannot be undone.

Do you work with partners?

Not yet, but we are keen to work with people who want to help test and improve JADE. Get in touch.