A teleprompter for authors — read your book without looking down

You wrote ninety thousand words; the marketing asks you to say two hundred of them into a lens. It is a different craft, and nobody warns novelists about it. The launch does not care: announcement video, excerpt readings, preorder pushes, event promotion — a fortnight that wants a dozen pieces of you on camera.

The reading itself has a built-in flaw. An author reading from the physical book gives the camera the top of a head — eyes down at the page, voice into the paper. The sentence you spent a morning on deserves eye contact, and the book in your hands makes eye contact impossible.

The formats multiply faster than the fortnight does. The bookshop event wants a thirty-second promo; the newsletter wants a two-minute reading; the retailer page wants audio; a podcast host wants a short clip to trail the episode with. Each is the same excerpt at a different length — the same performance problem four times over, on a deadline set by someone else’s printing schedule.

The podcast tour piles on. A book campaign now runs through other people’s shows, and the same six questions arrive on every one. The answers should be practised out loud — practised, not canned — and there is never a sensible way to rehearse alone.

And the writing and the recording happen on different devices. The excerpt gets tightened on the tablet at night; the phone in the morning has the old version; somebody records the wrong draft.

Reading from the physical book, eyes down, is the standard author video — and the least watchable format on any feed.

Memorising the excerpt is the alternative, and authors are the worst possible audience for it: you hear every deviation from your own sentence, and the self-monitoring shows on your face. Prose written to sound effortless comes out guarded.

Publishers help less than authors expect, too. A marketing department supplies graphics and a hashtag; the on-camera work is assumed, unbudgeted and yours. The debut novelist records it in the spare room after the day job — precisely the person and the hour least equipped for fifteen takes.

The rehearsal problem mostly goes unsolved. Answers get workshopped live, on the third podcast, where the good phrasing finally appears — and the first two shows never hear it. The scripts, meanwhile, are scattered across a notes app, an email draft and two files called excerpt-final.

Talk2Camera scrolls your excerpt under the lens as you read it. Voice follow runs on the device: pause at the line you always pause at, and it waits with you. The rhythm of the prose survives, and the camera gets your eyes for the whole reading instead of your parting.

The eye-contact analysis tells you whether that worked. After a take: how much of it you spent looking at the lens, your longest look away, how steady the framing was. The look-down habit of a lifetime of page readings becomes a number, and a number is fixable. When no face is found, it reports zero rather than guessing.

Audio-only takes cover the rest of the campaign. Same prompter, same pause behaviour, no video: record the excerpt as narration for people who ask what the book sounds like, and rehearse the podcast answers out loud — script the six questions you always get, read your answers to the prompter, listen back, tighten, go again.

One script serves every length of the campaign. The full excerpt is one script and the thirty-second cut is another beside it, read the same way, from the same place — so the reading you rehearsed once carries across the promo, the newsletter video and the narration, instead of being re-solved per format.

And the script follows you between devices. Tighten the excerpt on the tablet at midnight and it is on the phone at nine, reconciled by one shared ruleset on both platforms. Cloud script sync is part of a subscription — worth knowing before launch week rather than during it.

Two weeks before publication, a novelist blocks out one morning. Six scripts are ready — the announcement, two excerpt readings, three short thank-you-and-preorder pieces — written on the tablet over the weekend and waiting on the phone.

The first excerpt take reports 68% of it at the lens, longest look away four seconds: the page habit, visible as a number. Take two, with the prompter trusted, keeps nearly the whole reading at the lens — and the pause before the last line holds, because the script waited through it.

In the afternoon, audio-only: the opening pages read as narration for the website, and the six podcast answers recorded, played back and tightened — so the phrasing that used to surface on the third show is ready for the first. By evening, launch week is a folder of finished takes instead of a list of dreads.

And none of the morning needed a second person, which is the quiet point. A reading with eye contact used to mean a camera operator, or a studio teleprompter angled over a lens. Here the prompter, the camera and the coach are the same phone on a shelf — and the tablet the excerpt was edited on is already in step with it.