Speaker Notes (.docx)
Left drop zone / file picker (compact rows on mobile).
Turn notes and/or a slide deck into a polished A4 handout, then refine it in chat. The first result is a draft — AI can make mistakes, so review and revise before you print. Clearer preferences and chat instructions usually produce better results.
Layout of the upload screen (what to do is in Step by step).
Left drop zone / file picker (compact rows on mobile).
Right drop zone / file picker (compact rows on mobile).
Where the controls sit on the Preferences screen.
Content & Layout, Audience & Context, and Event & Branding panels.
Color palette and Theme tone. Footer: Back to Uploads and Generate Handouts.
Layout after generation starts (how to revise and export is in Step by step).
Revision conversation and controls on one side of the screen.
Print-size A4 pages of the handout on the other side.
The same two areas become Chat and Preview tabs. On desktop they usually sit side by side.
Add speaker notes as Word (.docx) and/or a slide deck as PowerPoint (.pptx). Either alone is enough; both is best when you have them. Max 50MB per file.
Privacy: uploads are deleted when the session ends or the page reloads.
After upload, configure how the handout should read and look, then generate. Preferences appear before generation (including when you arrive via Speaker Notes Generate → Handouts with notes preloaded).
Then choose Generate Handouts. You can still change colors and wording later in Revision Chat; preferences mainly steer the first draft. Clearer additional instructions and settings usually produce a better first pass.
Start generation and keep the tab open. Duration is not fixed: it depends on file size, image count, and current traffic. A realistic range is about 3–15 minutes for most runs (shorter for small, light files; toward the long end or beyond for large decks or heavy load). Progress moves through opening files, extracting media, understanding images when present, building structure, writing content, layout, and finalize.
Treat the result as a first draft, not a finished print-ready artifact. AI can misread material, omit points, or choose awkward layout — that is normal; plan to revise.
Use natural-language requests to fix wording, layout, colors, sections, and images. Be specific when you can (section name, “first image”, “make headings navy”). Clearer prompts usually produce better edits; vague asks are more likely to miss. See Revision chat below for what works and what does not.
On mobile, switch to the Preview tab after each revision (and after generate finishes) to confirm pages, images, and layout — Chat alone does not show the full print view. Switch back to Chat to send the next request.
Undo last revision reverts the previous successful chat edit if a change went wrong. Use it before stacking more revisions on top of a bad step.
When the draft looks right, use Convert & Export PDF (print flow). Preview zoom does not change the print size — export is at full print scale. Allow pop-ups if the browser blocks the print window.
Requests map to supported document operations. Expand a row for a short “how” note. Prefer one clear goal per message when possible. AI revisions can still miss — check Preview and refine with clearer instructions if needed.
Ask to change tone, fix wording, make a section shorter or longer, or rewrite a paragraph. Name the section or heading when you can so the edit stays focused.
Point at the heading title or quote a short snippet. Surgical text edits beat “fix everything” for predictable results.
Change colors for headings, body, or cover-related surfaces using plain language (for example “make headers navy”). Styling follows allowlisted document surfaces, not free-form design tools.
You can request larger or smaller type on body or headings, or fonts from the app’s supported set. Custom font uploads are not available.
Refer to an image by order (“first figure”) or description. You can remove it, change size, or swap it for a text stand-in. Photo cropping or recoloring the image itself is not supported.
Update cover title / event / date-style fields when present. Structural ops can move, insert, or delete sections when the coach can map your request to the outline.
Use the Undo control to step back one applied revision. This is the safety net after a chat change you do not want to keep.
Full-document translation is outside the supported revision path. Rewrite tone or simplify English wording instead.
Chat does not act as an image editor. Delete, resize, or replace an image with text; it will not crop or recolor the picture file itself.
Styling goes through allowlisted surfaces and tokens. You cannot inject arbitrary CSS or upload a font file.
Complex mixes of content and style may prompt a short plan or clarifying questions before applying. Split into clearer steps (wording first, then colors, then images) for fewer surprises.
AI can make mistakes. Generated handouts and chat edits can omit details, mis-tone content, or place images awkwardly. Always review before you print or share.
Draft first, polish second. Generation aims for a solid first pass; Revision Chat is where teaching tone and layout get personal.
Clearer prompts work better. “Shorten the Risk section” or “delete the second image” beats “make it better.” Specific goals and constraints usually produce better results.
Check Preview after changes. Especially on mobile, open Preview so you see pagination and images, not only the chat reply.
Export when done. Session work is not a permanent archive. Export PDF before you leave if you need a keepable file.
On smaller screens you switch between Chat (revision chat and controls) and Preview (print-size handout pages). After generate or a revision, open Preview to see the result. Pinch to zoom and scroll to pan.
Notes as .docx and/or a slide deck as .pptx. Either file alone is enough; both is fine. Max 50MB each.
Level of detail, strict file-content-only mode, audience and setting, event/branding fields (with optional AI extract), additional instructions, color palette, and writing theme (tone). Then generate handouts.
When Strict File Content Limit is on, the handout should only use information from your uploaded files — no external context.
No — generation produces a draft. AI can make mistakes. Use Chat to refine it, Undo if a change goes wrong, then export PDF. Clearer preferences and revision prompts usually produce better results.
Yes. Drafts may omit points, misread tone, or place images poorly; chat edits can miss too. Review in Preview before you export. Specific instructions (section names, what to change, what to keep) usually work better than vague asks.
There is no fixed time. It depends on file size, how many images are involved, and current traffic. A realistic range is about 3–15 minutes for most runs — shorter for small, light files; toward the long end or beyond for large decks or busy periods. Keep the tab open while it runs.
Wording and tone, shorten/expand, rename headings, colors and supported fonts, font size steps, image delete/resize/text stand-in, cover fields, and many section structure edits. See the collapsible list above for “how” notes.
Full translation, photo crop/recolor, free-form CSS, and custom font uploads. Prefer clear, single-purpose requests that name a section or image.
Your set of presentation slides — usually a PowerPoint (.pptx) file.