You're the best man (or maid of honor) at a wedding. You asked the couple's siblings to enrich their photos with KontexVoX — each of them dictated an anecdote while looking at each photo. Now you want a slideshow projected during your speech, where the anecdotes appear at the right moment, synchronized with each image.

A classic tribute slideshow is photos scrolling by to music. Everyone's seen that. What changes here: every photo carries an untold anecdote, told by someone who was there. The slideshow becomes a living story, not a silent scroll.

Prerequisites

Before you start, you need:

1. Photos enriched with KontexVoX. The descriptions live in each file's EXIF Description field. You can enrich your own photos, or ask the couple's relatives to do it — even better, it multiplies the points of view. See What is an EXIF file?.
2. An AI agent with file-system access: Claude Code, Codex CLI, or Gemini. Not a web chatbot.
3. The photos exported to your computer via AirDrop or the Files app (the channels that preserve EXIF — see How to share without losing the EXIF).
Your relatives' anecdotes are in the EXIF. A web chatbot erases them on upload. To read them, you need an agent that works on the files themselves.

What you've collected: an example

Say three relatives enriched their photos. Here's what the EXIF contains:

IMG_0312.jpg → "Pierre at 14, the holidays in Brittany. He always wore that old red beanie, even in August." (by his sister) IMG_0567.jpg → "The first time Marie met the family. She broke two glasses doing the dishes — we knew she was one of us." (by his mother) IMG_0891.jpg → "Their first apartment. The couch was so small they slept head-to-toe." (by his brother)

These anecdotes, your relatives dictated out loud. KontexVoX transcribed them and engraved them into the file. You now have raw material nobody has ever heard — and that's what will make your speech.

The prompt

Read all the JPEG images in this folder. For each image, extract the EXIF Description field and the capture date. Generate a full-screen HTML slideshow. Each photo appears in sequence with a crossfade. The anecdote extracted from the EXIF appears as an overlay at the bottom of the image, in an elegant serif typeface. Transition between photos on click or with an 8-second timer. Add a "speech" mode that displays the anecdotes large, for reading aloud.

What you get

A projectable full-screen HTML file for your speech. Photos advance on click — you control the pace. The anecdote displays large in "speech" mode, designed to be read aloud while the photo is on screen.

The effect on the room is immediate: you read the anecdote as the photo appears, and everyone discovers a story they didn't know, told by someone who isn't on stage. It's a tribute in several voices, orchestrated by a single slideshow.

The file is self-contained — put it on a USB stick, plug it into the projector, it opens in any browser. No internet connection, no account, no software.

Variation for another tool

With Codex CLI: also ask for a timing file (matching each photo to an estimated reading duration, to pace the speech). With Gemini: also ask for transition text between photos, to smooth out your delivery.

Frequently asked questions

Can I sort the photos in a specific order?

Yes. Add to the prompt: "Sort the photos in the following order: [list of file names in the desired order]." Without a sorting instruction, the agent will use the capture date — which works well for a chronological story, but not if you want a thematic order.

How many photos for a 5-minute speech?

One photo every 15-20 seconds on average, so 15-20 photos for 5 minutes. If you have more material, select the best before running the prompt — the quality of the anecdotes beats quantity.

Does it work if several people enriched the photos?

That's actually the ideal case. Descriptions from different relatives give varied perspectives — the sister tells the childhood, the mother tells the first meeting, the brother tells the recent stories. The slideshow becomes a chorus of tributes. Just make sure all the photos are in the same folder.

Can I do the same thing without KontexVoX?

Yes — you can type the anecdotes by hand into each photo, then run the prompt. But asking several relatives to type their anecdotes on a keyboard into their own photos is rarely realistic. Voice changes everything: each person tells their photo while looking at it, with zero typing effort. That's why I created KontexVoX.

Going further

Organizing a wedding album with anecdotes Generating a family story from contextualized photos Managing the copy album: understanding the enriched copy Check that your photos are properly enriched Optimize an enriched-photo session Memories page