Evidence Table Builder (App Maxxing, early release)

EARLY RELEASE: PLEASE READ BEFORE BUYING

• This app needs your own OpenAI API key. It will not run without one, and a ChatGPT subscription does not include one.
• It is an early release that still needs further testing. So far each app has passed our own offline tests, and has been run on its built-in sample against the live service on both models (October 2026). It has not yet been tried on a wide range of real files of different kinds and sizes. You may meet faults.
• Please tell us what you find after you buy: email kit376351@gmail.com with the subject "Maxxing feedback". Every report is read, and fixes are sent to buyers free.
• If we cannot get it running for you, email us within 14 days for a refund of the app price.

Build one table from papers you may send: one row a paper, each value with its sentence and page, sent in planned batches.

WHAT IT DOES

Add 3 to 40 papers and reports you already hold and are allowed to send, as PDF files with a text layer (ones whose text you can select and copy), Word files or text files, and choose the columns of the table: keep the starter set, rename a column, or add your own. Before anything is sent, the app plans the job on your computer. It sizes every paper, puts the papers into requests ("batches") that stay under the size where OpenAI doubles its price, works out the pauses needed to stay under your account's limit for each minute, and shows you the plan with an estimate for each batch. A paper that is too long for one request is left out, with a plain message saying how many pages would fit. The model then reads the papers of each batch and returns, for each paper and each column, what the paper states, the exact sentence and the page, or "not stated". Everything after that is done by the app itself, in code. It looks up every sentence in the paper it is said to come from and shows the page, or the heading in a Word or text file. It looks up every figure of a cell in that sentence. A cell whose sentence or figure it cannot find is shown as "not found in the file", not as a value. In a column that holds a number with a unit, it brings the figures to one unit and shows the conversion (5600 psi = 38.61 MPa); a figure in a unit it does not know is marked and left as written. It marks a paper that gives two different figures for one thing, such as an abstract and a results section that disagree. You choose where the job stops: tell the app to stop after any batch, and the finished rows are kept; the next run sends only the papers that are not finished. If OpenAI or your spending limit for one run stops a job part-way, the batches that finished are kept as well. You get a count of what was checked and what was not, the table, every sentence with its page, an Excel file and a Word summary. For educational purposes only. For your own research and your own documents only. Only add files you wrote or have the right to use this way.

HOW IT IS BUILT

• The app's own code does the finding, the arithmetic and the unit conversions. The AI is used only for what code cannot do, and what it says is checked against your file.
• Anything the app cannot find in your file is marked "not found in the file", not trusted.
• The cost is shown before each run and after it, with a limit you set for one run.
• No verdicts: the app never says pass, fail or compliant. It shows you what it found and where.

WHAT YOU NEED

• Windows, Mac or Linux, and free Node.js 18 or newer
• Your own OpenAI account with API access and a key (platform.openai.com). This is separate from a ChatGPT subscription.
• You pay OpenAI for what you use, on top of the price of the app. In our own tests on a short sample file, a run of this app cost about 35 cents. Longer files cost more.
• The app uses OpenAI's GPT-6 Astra model. In Settings you can switch to a cheaper model at about a fifth of the price.
• About ten minutes of setup. A step-by-step guide is included.

HONEST LIMITS

The app never finds or adds a paper: it reads only the files you give it. It is for your own papers and reports, open-access papers, and papers whose licence allows you to send them to an AI service; many publishers do not allow it, so check first. It does not store, republish or share your files, and the sentences it quotes are short and shown only to you. A sentence that really is in the paper can still be misread, and "not stated" does not prove that a paper is silent: the model can miss a statement, and the app cannot check a gap. Values that a paper gives only in a figure, a graph or a scanned table cannot be read or checked, and a paper with no text layer is marked "not checked (scan or image)". In a PDF set in columns, a sentence that runs from the foot of one column to the top of the next may not be found. The app converts units; it does not judge whether two studies are comparable, it never ranks the papers and it gives no judgement of a paper's quality. The plan uses an estimate of 1,300 tokens a PDF page: "Check the cost" measures the real size before you run. The sentences quoted from your papers are shown as written, whatever words they use. If you press Stop while a job is running, the batches that had finished are kept, and OpenAI still charges for the batch that was cut off. Finished rows are kept only while the app is open; to carry on another day, keep the Excel file and add it with the papers. Published standards are not an input. This is not a substitute for a second person extracting the data, and not a systematic review. A full table does not mean every cell is right.

For educational purposes only, and for your own research and your own documents. The app points things out for you to check: it does not certify, approve or sign off anything, what the AI writes can be wrong, and it is not a substitute for review by a qualified person. You are asked to accept this when you first open the app.

PRIVACY

The app runs on your computer. It says before each run exactly what will be sent to OpenAI, under OpenAI's terms, and what stays on your computer; nothing is sent to us. Use it only with documents you are allowed to send to an AI service: not an employer's or customer's confidential documents, not export-controlled material, and not someone else's unpublished work. Your key is stored only on your computer and is sent only to OpenAI.

WHAT IS INCLUDED

The app (a zip file), start files for Windows, Mac and Linux, a sample to try, a setup guide and a README. Kitson Workshop is not affiliated with or endorsed by OpenAI.

REFUNDS

Digital download. If we cannot get it running for you, email kit376351@gmail.com within 14 days for a refund of the app price. Credit you bought from OpenAI is theirs to refund or not.

FEEDBACK AND IDEAS

Email kit376351@gmail.com with the subject "Maxxing feedback" to report a fault or tell us what worked. For a new app idea, use the subject "Kitson idea".