The problem
The failure is not that making a deck is hard. It is that the same content has to exist in several places at once, and nothing keeps those places in agreement.
What actually happens
A team ships a feature. The facts — what it does, for whom, the three numbers that matter — end up in a slide deck for the launch meeting, a short video for social media, a page of documentation, a paragraph in the quarterly report, and a line on the public roadmap. Five artefacts, five tools, five copies.
Then one number changes.
In the best case someone remembers four of the five. In the normal case the deck is updated because it is presented on Thursday, and the others keep the old number until someone notices. Nobody decided to publish something false; the shape of the work made it inevitable.
Why the usual answers do not hold
| Answer | Why it falls short |
|---|---|
| "Export from one tool to another" | An export is a copy with a timestamp. It is correct exactly once, and nothing tells you when it stopped being correct |
| "Keep a single source of truth doc" | The doc is not what gets presented. The deck is rebuilt by hand from it, which is the copy again, one step further away |
| "Use a headless CMS" | It solves storage, not rendering: you still write the deck somewhere else, and the CMS cannot be read in a pull request or diffed |
| "Write slides in Markdown" | Closest, and where Récit starts — but it stops at one format. The film, the report and the book are still separate tools |
What the content deserves
Three properties, which together are rarer than they sound:
- Readable by a human in a text editor. Not a proprietary file, not a database row. Something you can grep, diff, review in a pull request, and still read in ten years.
- Typed for a machine. An agent, a CMS or a script should be able to produce and validate it without parsing prose. That is why every story has a JSON twin validated by a schema: same content, same structure, convertible both ways and tested in both directions.
- Independent of the renderer. The story must outlive the tool that renders it. If Récit
disappears, a
.mdfile is still a.mdfile.
Who this is for
Someone who already writes in Markdown and is tired of leaving it to make the slides. A team whose documentation, release notes and launch deck repeat each other. Anyone who has to publish the same thing in two languages and has watched the second one fall behind.
Not for someone who wants a drag-and-drop editor. That is a real need, and Récit is the wrong answer to it — see what it is not.