DEV Community: pierre

From Markdown to Explainer Video: A Developer's Workflow for Documentation People Actually Watch

pierre — Thu, 15 Oct 2026 07:05:49 +0000

If you write documentation, you’ve seen the analytics: your carefully crafted docs get opened, skimmed for eight seconds, and abandoned. Meanwhile the video explaining a tenth of what your docs cover gets all the engagement.
This isn’t a rant about users. It’s physics: reading is high-friction. A reader has to maintain their own context, supply their own pacing, and stay focused with no sensory support. A viewer gets pacing, structure, and redundancy handed to them — visuals, voiceover, and subtitles carrying the same idea on three channels at once.
The problem is that for developers, the gap between “I have great docs” and “I have a video” has historically required a whole production stack: scriptwriting for the ear, storyboarding, recording, editing, subtitling. Nobody sprints through that for internal documentation.
I’ve been testing a workflow that collapses that gap: NoteGPT’s AI Explainer Video Maker. This post is the practical rundown — what the workflow looks like, where it fits in a dev context, and where it doesn’t.
What It Actually Does (No Magic)
You give it a PDF, a document, or plain text. It returns an explainer video — visuals, AI voiceover, auto-generated subtitles — plus an exportable text script.
The important nuance: this is content-to-explainer, not generative text-to-video. The source of truth is material you’ve already written. That distinction matters for us, because:
Your docs (presumably) went through review. The facts are checked.
The tool’s job is translation — from “written to be read” into “organized to be listened to and watched” — not invention.
You keep an editable artifact at every step, which is the difference between a workflow and a slot machine.
The Workflow
Step 1 — Feed it source material
input: README.md / API guide section / onboarding doc / PDF / pasted notes
output: draft narration script
The AI first converts your material into a narration script: what gets said first, what comes second, where the emphasis lands. Written logic ≠ spoken logic — this conversion is the genuinely hard part, and it’s the step the tool invests in.
Step 2 — Edit the script before rendering
This is the part I want to dwell on, because it’s the difference between “AI slop” and a usable pipeline. The script is generated as a draft you review and edit before anything renders. Wrong emphasis on step 3 of your workflow? Fix the line. Missing a caveat? Add it. You stay the reviewer; the AI is the first-pass writer.
In dev terms: think of it as a PR you must approve before merge, not code that deploys straight to prod.
Step 3 — Pick a visual style
The same content can wear different visual languages. The samples show the range:
A solar-system explainer in a warm, curiosity-driven editorial style — watch
An “how does the internet work” explainer in a modular, block-frame style that suits step-by-step technical content — watch
A dopamine-and-motivation explainer in a bold poster style — watch
A moon-landing explainer in a broadside style that turns historical material into a staged story — watch
The pattern to notice as a technical viewer: visuals appear in sync with the narration. When the script introduces a concept, the visual for that concept shows up — the viewer never reconstructs structure from prose alone. That sync is what makes an explainer different from slides with a voiceover.
Step 4 — Pick voice and language
Voice carries tone: an internal SOP walkthrough and a public product explainer shouldn’t sound identical. Multi-language support also means one piece of documentation can serve international teammates without you recording anything.
Step 5 — Export MP4 + script
You get the video and the text script. The script matters long-term: your video becomes a repeatable, version-controlled format. When the docs change, you edit the script and regenerate, instead of re-cutting a timeline.
Where This Fits in a Dev Org
Some concrete use cases I’d route through this workflow:
Use case
Why it beats the status quo
Onboarding docs
New hires get a watchable walkthrough of your systems in week one, revisit it anytime, subtitles included for silent viewing
Architecture explainers
Turn that design doc nobody reads into a 3-minute video for the next team touching the service
SOP / runbook summaries
The “watch this before your first deploy” artifact that currently doesn’t exist
Open-source project READMEs
A demo video in the README that actually gets made, because the cost is one edit pass instead of a production weekend
Conference/meeting prep
Turn your notes into the explainer you were going to awkwardly narrate live anyway
Honest Limitations
No unverified hype — here’s what this is not:
Not a cinematic generator. If you need brand films or pixel-level art direction, use a production pipeline.
Not a replacement for good writing. If the source material is confused, the explainer will be confused with visuals and a voiceover. Garbage in, narrated garbage out.
Not a reason to delete docs. It’s a lower-friction entry point — the video is the front door, the docs remain the reference.
TL;DR
Reading is high-friction; watching is low-friction. Your docs deserve both formats.
The workflow: material in → editable AI-drafted script → pick style/voice/language → export MP4 + script.
The killer feature for devs is that the script is reviewable before rendering — it behaves like a PR, not a slot machine.
Try it with one small, familiar doc first: https://notegpt.io/explainer-video-maker
Turn your knowledge into videos that explain.

AI Deck Tools Keep Making the Same Mistake — And It's Costing You Hours

pierre — Thu, 17 Sep 2026 09:03:38 +0000

Every AI presentation tool I've tried in the last two years makes the same mistake: it generates first, and lets you discover structural problems second.
The 20-slide trap
Here's the cycle. You feed in your content, hit "generate," and get 20 slides in a minute. Then you read through and realize the argument doesn't flow — slide 4 should be slide 12, the conclusion needs to move up, and two sections are redundant.
Now you're reworking a finished deck instead of editing a plan. That's the expensive kind of rework, and it's built into how most tools work.
The fix: outline before slides
PPT Maker changes the order. It reads your content and proposes a structured outline first. You confirm it — change bullet points, add or remove sections, reorder — before a single slide is generated.
Editing a five-line outline is cheap. Editing twenty finished slides is not. It's the same reason writers outline before drafting, and it's wild that AI deck tools ignored it for this long.
The second half: single-slide edits
The other recurring pain is revision. You finish a deck, send it around, and someone says "just change the pricing slide." In most tools that means regenerating the whole thing and hoping it still looks right.
PPT Maker lets you change one slide with a prompt. The rest of the deck stays exactly as it was. For anyone whose deck goes through three revision rounds before a meeting, this is the feature that actually matters.
Two modes for the same content
PPT Maker splits generation into two modes:

Agent — structures your content into a clear, logical narrative. Good for reports, research updates, training material.
Creative — turns key pages into more visual, designed spreads. Good for proposals, launches, brand-facing decks. Same input, two different expressions. You can use Agent for the structure and Creative for the pages that need impact. Input and output You can start from a topic, pasted text, a PDF, an article link, or a video. Export to PPTX (which drops into any existing workflow), or PDF, Keynote, PNGs. You can also download just the outline. What it's not To be fair: this is an efficiency tool, not a replacement for a professional designer. The visual style is clean and restrained, not flashy — no 3D cards or motion effects like Gamma. Chart customization is basic. Very long inputs (like a 30-page report) can produce duplicate section headings you'll need to merge manually. It's browser-only, no offline client. If you need maximum visual polish, Gamma or a human designer is still the answer. If you need to get from "a pile of material" to "a deck I can actually present" with minimum rework, that's where PPT Maker wins. The bottom line The bottleneck in deck-making was never "I can't draw slides fast enough." It was structure and revision. PPT Maker is one of the first tools built around those two problems, and it shows. Free tier is 50 credits a month, enough to try it properly. Give it a real document and watch the outline step save you the rework you didn't know you were doing. https://pptmaker.io

I Stopped Using AI for Individual Steps and Started Giving It the Whole Task

pierre — Thu, 03 Sep 2026 02:46:22 +0000

Most of my AI workflows have been surprisingly manual.

Ask ChatGPT something.

Copy the answer.

Open another tool.

Search for additional information.

Paste everything into a document.

Ask AI to restructure it.

Repeat.

AI made individual steps faster, but I was still the workflow engine connecting everything together.

Recently I tried a different approach with NoteGPT’s AI Agent.

Instead of asking it to complete one step, I gave it the end goal.

For example:

Research five competing products, find their recent updates, compare their features and positioning, and produce a structured report.

The agent breaks that goal into smaller tasks, decides which tools it needs, browses the web, processes the information, and combines the results into a final deliverable.

That’s a subtle but important difference.

Chat vs. delegation

A normal AI conversation usually looks like this:

User → prompt → response → next prompt → response → next prompt.

An agent workflow looks closer to:

User → goal → plan → tools → execution → result.

You’re moving some of the orchestration away from yourself.

I found this especially useful for tasks involving several different operations: research, summarization, comparison, writing and formatting.

No agent framework required

The other thing I appreciated was the lack of setup.

I’ve experimented with agent frameworks before, and many of them quickly turn into a project of their own.

Environment variables.

API keys.

Model configuration.

Tool definitions.

Workflow logic.

Debugging.

That flexibility is great when you need it, but sometimes I just want to delegate a task.

NoteGPT takes the opposite approach. It runs in the browser, requires no code or API key, and lets you switch between 20+ models including GPT, Claude, Gemini and DeepSeek.

There are also predefined skills for common workflows such as Deep Research, AI Documents, AI Slides, AI Sheets and Data Analysis.

Where I think this is useful

I wouldn’t hand an agent an important task and blindly trust the output.

Research still needs verification. Code still needs testing. Anything involving financial, legal or medical decisions obviously needs proper human review.

But there’s a large category of boring work where the cost isn’t intelligence — it’s coordination.

Finding information.

Moving information.

Restructuring information.

Turning it into another format.

That’s where agents are starting to make sense to me.

Instead of making each individual step 30% faster, they can sometimes remove several steps entirely.

NoteGPT’s AI Agent is free to start if you want to experiment with this workflow:

notegpt.io/ai-agent

The question I’m increasingly asking myself isn’t “Can AI answer this?”

It’s:

“Can I stop doing this workflow manually?”