Help
Last updated: 15 November 2026
This page explains what each part of the cue card converter does and what to do when something goes wrong. Everything described here is how the tool behaves today. If your question isn’t answered, contact us.
Adding your text
- Paste: open “I’d rather paste the text”, paste your speech and choose “Convert to cue cards”.
- Upload a file: choose a file or drop it on the upload area. Files up to 40 MB are accepted.
| File | What is read |
|---|---|
| Word (.docx) | The document text. Older .doc files are not supported: open them in Word and save as .docx first. |
| The selectable text, with paragraphs rebuilt from the layout, hyphenated line breaks joined and lines that only hold a page number removed. A scanned PDF has no selectable text, so nothing can be read; paste the text instead. | |
| PowerPoint (.pptx) | The speaker notes, slide by slide in presentation order; each slide starts a new section headed “SLIDE n”. If no slide has notes, the slide text is used. Hidden slides are skipped. |
| OpenDocument (.odt) | The paragraphs and headings. |
| TXT, Markdown (.md) | The text. In Markdown, code blocks and images are removed, links keep only their words and list items become bullets. |
Files are opened in your browser; only the extracted plain text is used for the next step.
How the text becomes cards
The text is divided into parts: each timed section (for example “19:30 – WELCOME”) is one part, otherwise paragraphs are grouped into parts of about 450 words. Up to six parts are analysed at the same time, so the parts of a long script are analysed in parallel. For each part the AI chooses where to break the cards and marks names, titles, organisations, numbers, dates, times and bracketed stage cues such as [MUSIC].
The AI is not allowed to rewrite you. The converter compares its result with your original text, and if any word, number or punctuation mark was changed, that result is thrown away. Your text is then split on your device at section and sentence ends, without the AI’s highlights: numbers, dates and times are still highlighted automatically, but names, titles, organisations and stage cues are not. A notice tells you so. The same happens if the AI service doesn’t answer within 45 seconds or you are offline. Short neighbouring sections are put on one card when they fit at the card’s real size.
Card size, paper and words per card
| Card size | Measurements | Default words per card (English) |
|---|---|---|
| A5 | 148 × 210 mm | 100 |
| Half Letter | 5.5 × 8.5 in | 100 |
| A6 | 105 × 148 mm | 60 |
| 4 × 6 in index card | 4 × 6 in | 60 |
| 3 × 5 in index card | 3 × 5 in | 40 |
Paper is A4 or US Letter, and cards can be landscape or portrait. The defaults follow the region chosen in the language and region menu. A card size you pick yourself is kept; choosing another region in the menu changes the paper size. If you change the card size after the cards were made, the text is shrunk to fit; choose “Re-split cards for this size” for cards at a readable size. “Words per card” applies to new uploads and revisions. How these numbers were measured is in the cue card template guide.
Editing cards
- Click inside a card to change the text.
- Highlight / Remove: select text and highlight it, or click inside a highlight to remove it. Under “Show highlights for” you can hide a whole group (names, titles and organisations; numbers, dates and times; stage cues; key phrases).
- Split here splits the card at the cursor; each card also has buttons to move it up or down, merge it with the next card, add a card or delete it.
- A card that is too full is marked “overflows, split it!”, and a card that ends in the middle of a sentence is marked too.
- Undo steps back through card edits (up to 50 steps).
- Load revised text: upload the updated file and the cards are made again. Your design settings stay and matching manual highlight choices are applied again, but edits typed into cards are not carried over, so copy them into the file first. Undo brings back the previous cards. More on this in editing cue cards after last-minute changes.
Printing, PDF and the back side
- Layout: “One card per page” (the card fills an A5 or card-sized page) or “Several cards per sheet (cut lines)” on A4 or Letter.
- Print opens the print file; PDF, PPTX and DOCX download the cards. Keep margins on “Default” or “None” and turn off headers and footers in the print dialog.
- For one card per page, choose the card’s paper size in the print dialog yourself: browsers don’t let a website change the printer’s paper size.
- Back side: “Print back side (duplex)” is off by default. When it is on, each card gets a back with an image you choose (or the default logo), the event name and a background colour, and you can set the image size. Print double-sided so each back lands behind its card.
Paper, duplex flip and scaling problems are covered step by step in how to print cue cards on A4 or Letter.
Practice mode and teleprompter
Practice shows one card at a time with an estimated time for that card and for the whole deck, worked out at your speaking pace (130 words a minute by default; change it under “Speaking time”). Move with the arrow keys, Page Up/Down, Space, a presentation clicker or a swipe; Escape closes it. There is a timer you can start and pause, and the screen is kept on where the browser allows it.
Teleprompter sends the plain text of your cards to the online teleprompter with the same pace.
What is saved, and what the AI receives
- Your cards and settings are saved in this browser (local storage), so “Continue where you left off” works after you close the tab. There is no account; clearing the site’s data in the browser removes them.
- To make the cards, the extracted plain text and the words-per-card number are sent over an encrypted connection to our server function and to our AI provider, Anthropic, which returns the card breaks and highlight marks. The file itself, its images and your design settings are not sent. We don’t store your text on our servers.
- The log entry our server writes for each analysis never contains the text or the cards, only technical figures such as the number of characters and cards and how long the analysis took. Hosting logs (IP address, browser, time) are described in the privacy policy.
- Once the cards exist, editing and printing keep working without a connection.
The full details are in the privacy policy.
Common problems
| What you see | What to do |
|---|---|
| “This file type isn’t supported” | Use .docx, PDF, .pptx, .odt, .txt or .md, or paste the text. Save .doc files as .docx. |
| “This PDF has no selectable text” | The PDF is probably a scan. Copy the text from the original document and paste it. |
| “The file could not be read” | Open the file and save it again (in Word: Save As → .docx), or paste the text. |
| PowerPoint text is not what you expected | Only speaker notes are used when any slide has notes. Add notes to the slides you need, or paste the text. |
| Cards show only number, date and time highlights and a notice about AI analysis | The AI didn’t answer, changed the wording or you were offline, so the text was split on your device. Highlight by hand, or try again with “Load revised text”. |
| “Too many requests in a short time” | Wait a few minutes and try again. |
| A card says “overflows” | Put the cursor where the card should break and choose “Split here”, or lower the font size. |
| Printed cards are the wrong size | Check the paper size in the print dialog and set scaling to 100% or “Actual size”. |
Frequently asked questions
Do I need an account?
No. The converter is free and works without an account. Your cards stay in your browser.
Can the AI change my words?
No. The result is compared with your original text, and if a word, number or punctuation mark was changed it is discarded and the text is split on your device instead.
Why is my scanned PDF empty?
A scanned PDF is a picture of text, so there is no text to read. Copy the text from the original document and paste it.
Can I use it offline?
Once the cards exist, editing and printing keep working offline. Making new cards with the AI’s highlights needs a connection; offline, the text is split on your device and only numbers, dates and times are highlighted.