Hundreds of family photos, each enriched by voice. You went through the old photos with your parents, and for each one you said out loud what you saw: who's there, when it was, what happened that day. KontexVoX transcribed and engraved each anecdote into the EXIF. Now you want to turn this collection of fragments into a story — a narrative text that tells the family's history through the anecdotes, like a journal.
A photo album is images with dates. A family story is a story. The difference is the connective tissue: why this photo follows that one, what happened in between, what the person in the photo was feeling. If every photo carries a description, the AI agent can weave those links and produce the narrative.
Prerequisites
Before you start, you need:
Description field. See What is an EXIF file?.What you dictated: an example
Here's what the EXIF of four family photos contains:
These anecdotes, you dictated while looking at the photos with your parents. KontexVoX transcribed and engraved them along with the capture date. The AI agent will sort chronologically and weave a narrative that ties the anecdotes together.
The prompt
What you get
A Markdown file containing a continuous family narrative. The anecdotes you told in front of each photo are woven into a text spanning years, even decades. It's a family journal, generated from your own words.
The narrative follows the chronology. The AI introduces natural transitions between episodes ("Two years later, on Marie's first day of school…"), without inventing anecdotes: everything comes from your descriptions. The tone is personal and warm, as the prompt requests.
The file is in Markdown, readable as plain text and exportable to PDF. You can print it, share it with your family, or keep it as a family document.
Recommended tool: Gemini
Gemini excels at long-form, multilingual text generation. For a family story that can run several pages, Gemini is better suited than Claude Code for this specific task. Use Gemini via AI Studio or the API.
Frequently asked questions
Does the AI invent anecdotes?
No. The prompt asks it to weave in the existing anecdotes, not invent new ones. The AI only connects your descriptions with transitions. Everything in the story comes from what you dictated. If you want the AI to elaborate more, add to the prompt: "You may expand the transitions, but invent no facts."
Does it work with photos spanning several decades?
That's the ideal case. Photos from 1990 to 2020, enriched with your parents' anecdotes, produce a story covering 30 years. The AI structures the narrative by era, following the chronology of the capture dates.
Can I aim for a different tone?
Yes. The prompt asks for a "personal and warm" tone. If you want something more factual (documentary) or more literary, adjust the prompt accordingly. The AI adapts.
Can I do the same thing without KontexVoX?
Yes — you can write the anecdotes by hand under each photo, then ask the AI to connect them. But 200 photos means 200 descriptions to type. Voice changes everything: you look at the photo with your parents, you say what you see, the anecdote is engraved. That's why I created KontexVoX: so the story feeds on the voice of those who were there, not on typing after the fact.