Notre vision.
Pas « comment utiliser Plain » — « comment nous voyons le Bureau à l’ère de l’IA ». Décisions produit, impasses testées, choix validés.
- No.012026-06-1910 min
Open-source AI presentation tools in 2026: the honest landscape
Presenton, ALLWEONE, Slidev, Marp, and the web-first path. A clear-eyed look at the self-hosted and open-source ways to generate slides with AI, and how to choose.
Lire - No.022026-06-189 min
How to turn a CSV into a dashboard with AI
You have a CSV of numbers and need a dashboard, not a 5,000-row grid. Here are the real ways to get there with AI, and how to make one that stays a living link.
Lire - No.032026-06-189 min
How to write a memo with AI that people actually read
A decision memo is a recommendation with its reasoning attached. Here is a paste-ready structure, the AI drafting loop, and the two things you must fix by hand.
Lire - No.042026-06-189 min
How to present from the browser (no install, no app, just a link)
You don't need PowerPoint or Keynote installed to give a presentation. Here's how browser-based slideshows work, what they fix, and when a link beats a file.
Lire - No.052026-06-179 min
How to turn Markdown into Google Slides (and when not to)
Three real ways to convert a Markdown file into Google Slides, what each one loses, and why a web-first link is often the better destination than a slide file.
Lire - No.062026-06-147 min
Markdown to web slides: Marp, Presenterm, MkSlides, and the AI + link option
The open-source CLI tools turn Markdown into slides for free — you install and host them. Here's how they compare, and when AI generation plus a shareable link wins.
Lire - No.072026-06-109 min
You can build a Dune-style dashboard without a chain indexer or a warehouse
Dune's value isn't the data pipeline — it's the loop where a SQL query becomes a shareable panel. That loop runs on your own CSV, in the browser.
Lire - No.082026-06-109 min
Google Docs bolted AI onto a page; the better move is AI that owns the source
A suggestion box inside a WYSIWYG doc can nudge a paragraph. AI that writes the Markdown source can rebuild the whole document — and re-render it as a doc, a deck, or a dashboard.
Lire - No.092026-06-109 min
For long-form writing, a Notion doc fights you; Markdown gets out of the way
Notion's block model is built for wikis and relational databases. For one author writing one 4000-word argument, every paragraph becomes a draggable object — and that structural friction is the real cost.
Lire - No.102026-06-109 min
Running SQL in the browser means your dashboard has no backend to break
How in-browser SQL engines actually work, what they can load, where they fall over — and paste-ready queries that run entirely client-side.
Lire - No.112026-06-1010 min
A Q3 strategy doc written in Markdown beats one trapped in a slide template
Strategy is an argument, not a bullet list. Write the reasoning first as one Markdown file, then render it as a doc for readers and a deck for the room.
Lire - No.122026-06-109 min
If you can't diff a slide, you can't trust the AI that changed it
Binary decks make every AI edit a black box. Text-source slides turn each change into a reviewable diff you can revert line by line.
Lire - No.132026-06-078 min
How to make a pitch deck from one Markdown file (and ship it as a link)
Markdown as source. Web as render. One source generates the deck, a backup .pptx, and a sharable URL.
Lire - No.142026-06-0711 min
Picking an AI deck tool in 2026 — Tome, Beautiful.ai, Pitch, Gamma, Plain
Honest selection guide. What each tool is actually for, who they're for, and where they break.
Lire - No.152026-06-069 min
Plain vs Gamma in 2026 — an honest comparison from someone who used both
I built decks in Gamma for three months. Then I built Plain. Here's how to pick — without the marketing spin.
Lire - No.162026-05-265 min
The document is a link, not a file
Why the right unit of delivery for AI-era work is a URL — and what changes when you stop attaching .pptx, .docx, and .xlsx to email.
Lire - No.172026-05-266 min
What AI deck tools got right in 2026 — and what they're still missing
The first wave of AI presentation tools nailed one half of the problem. Here's what we think the second half looks like.
Lire - No.182026-05-265 min
Eight design philosophies, and why AI shouldn't be allowed to mix them
Why we don't ship a theme picker — and what we ship instead.
Lire - No.192026-05-264 min
Plug Plain into Claude Code in 30 seconds
How to give your coding agent the ability to ship decks, docs, and dashboards — without leaving the terminal.
Lire - No.202026-05-264 min
Office files belong in the safety net, not the workflow
Why we still export to .pptx, .docx, and .xlsx — but only when you really need to.
Lire - No.212026-05-264 min
Building dashboards without a data warehouse
Most dashboards don't need ClickHouse, Snowflake, or BigQuery. Here's the case for keeping it small, fast, and on your laptop.
Lire - No.222026-05-265 min
From Excel to dashboard: the case for narrative data
Most spreadsheets aren't records — they're arguments. Here's why narrative dashboards beat grids.
Lire - No.232026-05-264 min
Diffable documents: why every artifact in Plain is git-ready
What happens when documents become version-controllable like code.
Lire - No.242026-05-106 min
HTML is for reading, not for source
trq212's thesis — HTML as the output format for agents — got 5.4M views on X. We agree with half of it and push back on the other half. That other half is the entire reason Plain exists.
Lire
24 articles · pas de rythme fixe