Markdown to Video for technical knowledge beyond the repository
Bring the Markdown source your technical team already maintains. Preserve the code, steps, and terminology while turning it into a narrated tutorial or product update for a broader audience.
Four steps from a Markdown file to a video people finish.
No timeline, no scene-by-scene scripting, no syntax to clean up. Paste the .md, review the plan containing the script, hit render.
01
Drop the Markdown in
Paste your raw Markdown or point at a public README URL, and ngram parses # heading levels, bullet and numbered lists, code fences, and blockquotes. Direct .md file upload is coming soon; paste and public URL work today.
02
Syntax becomes story
Each ## heading becomes a scene. Numbered steps build as sequence cards. Bullet lists animate in. Code fences render as readable on-screen blocks. The agent writes the voiceover from the prose, not the markup.
03
Review the plan containing the script
Approve, reorder, or rewrite scenes in plain language. Swap a visual, tighten a line, drop a section that reads better as text. Every change is one chat turn.
04
Ship and share
Export MP4 in 16:9, 1:1, or 9:16. Publish to a hosted /watch page, push to LinkedIn, or drop the link in the repo's release notes.
Output controls
Sensible defaults. Real controls when the file deserves them.
Heading-aware scene breaks
H1, H2, and H3 in the Markdown become real scene boundaries, not arbitrary 7-second chunks. A long section under one heading splits on its natural paragraph beats.
Code fences stay legible
Fenced code blocks and inline snippets render as monospace on-screen cards that hold long enough to read. The voiceover narrates the explanation around the snippet, not the syntax itself.
Verbatim or rewritten script
Keep the file's exact wording for a precise install guide, or let the agent rewrite dense prose for spoken voiceover. Toggle per scene.
Voiceover that fits the doc's tone
Brisk for a changelog, measured for an API reference, warm for a getting-started guide. Pick a brand voice or use a cloned one from /app/settings/voice.
Burned-in captions
Auto-generated and styled per brand kit. Numbered steps and command names stay readable; blockquotes hold for emphasis when the video plays muted.
Motion graphics for lists and tables
Bullet hierarchies animate in. Numbered steps stack as counter cards. Markdown tables render as branded scene cards instead of a flat grid of text.
Localized edits prepared separately
Render the same README as a German quickstart, a Spanish changelog, or a Portuguese integration guide. Captions, on-screen text, and voiceover all swap together.
Channel versions from one file
Prepare a 9:16 launch-post edit, a 1:1 LinkedIn edit, or a 16:9 hosted-docs edit from the same Markdown. Review each cut separately before exporting one MP4.
The rest of ngram
Markdown to video is one node. The product is the whole pipeline.
Source isn't Markdown? There's a converter for that.
A .md file is one starting point. Word docs, PDFs, slides, and plain text all route through the same agentic storyboard step and the same brand kit, so the output looks identical no matter which source you paste.
Fire Markdown to video from the repo, the editor, or the agent.
These recipes start from where your .md already lives: a merged README, a tagged CHANGELOG, a rendered docs page. Wire one up, or hit the REST API and self-managed webhook subscriptions for video.completed or video.failed to build your own.
REST APIMCP serverWebhook subscriptionsFor API automation, send supported text or URL inputs with a self-service API key, retrieve the video job status, and subscribe to video.completed or video.failed.
How it compares
If you've been doing this another way.
Slide-based Markdown tools render your .md as static deck frames. Avatar tools narrate the text over a presenter card. ngram reads the file's heading-and-list structure and builds a branded motion-graphic video with voiceover and captions.
Yes, headings become scenes, lists become sequences, code fences render legibly
Renders .md as static slides, no scene logic
Extracts text; avatar narrates linearly
Manual, depends on the editor
Time from .md to first cut
Minutes
Slides fast, then a manual screen-record pass
Tens of minutes (script + render)
Hours to days
Re-render when the file changes
Paste the new .md, re-render
Re-export slides, re-record
Re-generate narration
Re-edit from scratch
Brand kit across every render
Set up and select your Brand Kit
Theme CSS per deck
Template-level branding
Manual per project
Voiceover from the prose
AI voiceover written from the file, captions burned in
None, you record it
Avatar voiceover
Manual VO or none
Storyboard preview before render
Full scene plan, editable in chat
Slide preview only
Generate first, edit after
Editor decides
Localized variants
Voiceover, captions, on-screen text translate together
Manual per language
Avatar voiceover per language
Manual per language
Channel versions
16:9, 1:1, 9:16 from one file
Slide aspect only
One format per export
Manual reframe per format
Best when
You want a branded, scene-based video directly from a Markdown file
You want presentation slides from Markdown
You want a talking-head presenter for the text
You have an editor and time to burn
Common questions about Markdown to video
How do I convert Markdown to video?
Paste your raw Markdown into ngram, or point at a public README URL. The agent reads the # headings, bullet and numbered lists, and code fences, then builds a scene-by-scene storyboard. Review the plan in chat, approve, and the system renders a branded MP4 with voiceover and captions.
What Markdown elements does the converter understand?
Heading levels (# through ###), unordered and ordered lists, fenced code blocks and inline code, blockquotes, and tables. Headings become scene breaks, numbered lists become sequence cards, tables render as branded scene cards, and code fences appear as readable monospace blocks.
Can I convert a GitHub README to video?
Yes. Paste the raw Markdown directly, or share the public README URL and ngram fetches the rendered page. It parses the same heading structure, lists, and code blocks whether you paste the file or link to it.
What file formats does the markdown to video converter accept?
Markdown source as .md, .markdown, .mdx, and plain .txt, plus pasted Markdown text. File upload is rolling out; paste and public URL work today. Image-only or binary files are not supported here.
How long can the source Markdown be?
Up to about 15,000 words per render. Past that, split the file by top-level heading and render chapter videos. ngram can chain them into a single hosted playlist on a /watch page.
What does the final video look like?
A branded motion-graphic video with an on-screen scene card per heading, animated bullet and numbered lists, legible code blocks, AI voiceover, and burned-in captions. Brand kit colors, fonts, and logo apply to every scene.
How do I keep the video in sync when the Markdown changes?
Re-paste the updated .md and re-render. The storyboard rebuilds from the new structure in a few minutes, so a changelog or README video stays current without a re-recording cycle.
How is this different from Marp or a slide tool for Markdown?
Marp and similar tools render your Markdown as static presentation slides that you then screen-record yourself. ngram reads the file's structure and builds a motion-graphic video with AI voiceover and captions directly, no manual recording pass and no slide deck in between.
Can I translate the Markdown video into other languages?
Yes. The same source .md can render localized versions where voiceover, captions, and on-screen text all switch language together. Useful for global docs, multi-region changelogs, and developer quickstarts.
Does ngram offer enterprise controls for docs and developer teams?
Yes. Workspace brand kits, team roles and invites, and hosted /watch pages with link controls are available on paid and enterprise plans. Talk to sales about security, access controls, and data handling for your team.
Markdown → Video
Ship the video version of the Markdown file your team already wrote.
Paste the README, changelog, or docs .md, review the plan containing the script, and render a branded video people watch instead of bookmark.
Give customers a short explanation of what shipped, who it helps and where to find it. Use your release notes to make a repeatable video update, then check the result against the shipped product.