Talk2Camera
Menu
en-US
E-commerce sellers

A teleprompter for e-commerce product videos — pitch, then show

The problem

A product video has two halves that want different cameras. The pitch wants your face — why this bag, who it is for, what it costs. The demonstration wants the product — the stitching, the zip, the way the strap sits. And the moment you say “let me show you”, you hit the problem: stop recording, turn the phone round, start a second clip, and hope your energy matches when you stitch the two together later.

The specs make it worse. E-commerce buyers return things, and they return things because the video said the sleeve fits a 16-inch laptop and it does not. Fabric weight, battery life, internal dimensions — these have to be said exactly, and nobody holds forty SKUs’ worth of exact figures in their head.

Repetition compounds it. The same bag exists in four colours and two sizes; the shipping line changes with every promotion; the price moves for the sale weekend. Each re-record from memory drifts a little further from the listing copy, and by the fifth video of the day you are giving the tan bag the brown bag’s measurements. Nobody on set notices, because there is no set — the first person to notice is a customer, in a dispute, quoting your own video back at you.

Then there is the catalogue. One product video is a job; a range is a production schedule. Forty SKUs, each wanting its own words, and each wanting a different shape at the other end — vertical for the feeds, whatever the listing page takes, something else again for the channel.

Most sellers know all of this and film anyway, because the listing with a video outsells the listing without one. The question is not whether to shoot — it is how to shoot forty without losing a week.

How this gets made today

The common approach is memory: read the bullet points, put the phone on the shelf, and improvise. It produces the ramble — ninety seconds of talk for thirty seconds of information — and it produces rounded numbers. “About two hundred grams” is what memory does to 185 g.

The tidier approach is two shoots: pieces to camera in one batch, product close-ups in another, joined in an editor. It works, and it is slow — every join is a decision, the room tone changes between clips, and the edit for a single video is an evening. Multiply by the range.

And the scripts, where they exist, live in a notes app or a spreadsheet — one more screen to juggle between takes, and no answer, three weeks later, to the question of which take on the camera roll belonged to which SKU.

How it works with Talk2Camera

Talk2Camera puts the script under the lens of the phone that is filming. The specs are written once, checked once against the listing, and read exactly — gram for gram, port for port. You look where the buyer is looking.

When you reach “let me show you the stitching”, you flip to the rear camera mid-sentence. The recording does not stop — the encoder keeps running through the switch — so the pitch and the demonstration land in one continuous file. Flip back to close. There is no join to hide, because there is no join.

You choose how the flip looks: an instant cut, a 100 ms quick blend, or a 250 ms dissolve. The hard cut suits a punchy gadget video; the dissolve suits leather goods. Either way, the take never breaks.

The script library is what makes a range shootable in a day. One script per SKU, in a folder per collection, searchable, with a word count and an estimated read time on every one — so you know the video will run about forty seconds before you press record, and the shoot day runs script, take, next script, next take.

It also changes what a price change costs. When the strap length is revised or the sale price moves, you search the library, edit one line of one script, and re-shoot one forty-second video — not the range. The script is the single source the video is read from, so the correction happens in the words first and the footage second, and the two cannot disagree.

At the end, export presets sized for where the video is going: TikTok, Instagram, YouTube and LinkedIn, each with its own resolution cap and bitrate, the vertical feeds framed 9:16. The original file stays untouched for your product page or anywhere else. One honest note: on the free plan, exports carry a small Talk2Camera mark; the paid plans remove it.

A worked example

A seller with a leather goods range has nine new bags for autumn. Sunday evening: nine scripts, written from the listing copy, in a folder called Autumn. Each runs about eighty words — the library says thirty-five to forty seconds apiece.

Monday morning: phone on a stand at the end of the workbench. Script one under the lens. Pitch to camera, flip to the rear camera on “here is the grain up close”, a slow pan over the bag, flip back, close with the price. One continuous clip, no editing debt. Next script. All nine are shot before lunch.

Monday afternoon: each clip exported 9:16 for TikTok and Instagram, once more at the YouTube preset, and the untouched original uploaded to the product page. Nine listings get video in a day — and the figures in every one of them match the listing, because they were read, not remembered.

The features this uses