ivy.video
Tutorial · Substack Series

How to turn a Substack post into a video series

One long-form essay → a recurring per-subject video series. Stippled portraits, magazine-cover cards, kinetic typography. Each episode is ~75s, 4:5 for Instagram and LinkedIn feed.

Browse all prompts
75s · 4:5

why this format works

A Substack post is a list of ideas. A video series is a posting habit. Splitting one essay into a per-subject episode (one person, one company, one decision per video) gives you 5–10 weeks of distribution off a single piece of writing — and the format reads as ongoing curation rather than a one-off promo. Use this when you have a curated list inside a longer post (people you respect, companies to watch, books, papers) and want to reach an audience that won't read 3,000 words. Use a different format if your essay has one through-line and no natural splits — those should stay as a single video.

what ivy does

What's automated, scene by scene.

Reads the Substack post

Pulls the byline, the list of subjects, and the reason each one is in the list. If your post is paywalled, paste the text directly into chat — the scraper alone won't reach it.

Builds a per-episode template

Each episode shares a structural template: an intro card with you (the curator), a series title reveal, a hero card for the subject (photo + role + handle), the "why they're different" beat, three to five works as magazine-cover cards, and a CTA with the next-episode teaser. Variation lives in the subject, not the structure — so the series reads as one coherent thing.

Sources real portraits

For each subject, ivy fetches a Wikipedia portrait and stipples it (sepia/yellow paint-block treatment) so the look is editorial, not a stock LinkedIn headshot. If the subject isn't on Wikipedia, you can upload a photo and ivy applies the same treatment.

Writes the voiceover in series voice

Marcus (ElevenLabs, eleven_v3) by default — deep, deliberate, magazine-narration register. Per-beat MP3s rather than one continuous track, so individual lines can be re-recorded without re-rendering the visuals. Roughly 70–75s per episode.

Returns a brief you can edit before rendering

Re-order subjects, swap a portrait, change the "why" line, change voice — all in chat. When you say render, ivy locks the brief and produces an MP4. After episode 1 ships, subsequent episodes inherit the visual language so each one takes a fraction of the time.

make one in five minutes

Five steps. Five minutes.

01

Paste the Substack post URL or text

Drop the URL into chat at ivy.video. If it's paywalled, paste the body text. Tell ivy this is for a series — "split into one episode per person/company/item."

02

Confirm the subject list and order

Ivy proposes one episode per subject in the post. Re-order, drop, or add: "start with subject 3," "skip subject 5," "add an episode 0 introducing the series."

03

Render episode 1

Ivy returns episode 1 as an MP4 in 5-7 minutes (the first one is slowest because the visual language is being established). 4:5 by default for Instagram and LinkedIn feed.

04

Approve the look, then render the rest

Once episode 1 looks right, say "render the rest with the same visual language" — episodes 2-N inherit the templates and finish in 2-3 minutes each.

05

Post on a cadence

Don't drop all episodes at once. One per week (or whatever cadence your audience expects) keeps the series feeling alive and gives each subject room to be discussed in isolation.

what to watch out for

Hard-won lessons from shipping these.

List posts vs essays

This format works when the post has discrete subjects you can split on — a list of people, companies, books, decisions. It does not work when the post is a single argument with no natural splits. If you can't write the title of episode 1 in five words from the post alone, the post probably shouldn't be a series.

Portrait sourcing

Ivy uses Wikipedia first. If a subject isn't on Wikipedia, the brief will flag it and ask you to upload a photo. Don't let it generate a fake portrait — for a series like "thinkers I respect," a generated face reads as deeply weird. Real photo or no episode.

Aspect and length

4:5 (1080×1350) is the sweet spot for Instagram feed and LinkedIn. 75 seconds is long enough to land an argument and short enough to not lose retention. If you go vertical 9:16 instead, cap at 45-50s — feed reading habits are different.

Series voice consistency

Pick one voice for episode 1 and stick with it for the run. Switching narrator mid-series is jarring. If you want variety, vary the subjects, not the voice.

Magazine-card density

Each "work" or "thing they're known for" gets a cover-style card with a 5-word headline and a brief sub-line. More than 5 cards per episode and the pace tightens uncomfortably; fewer than 3 and the body of the episode feels thin.

Writing the series brief once

Lock the visual language (palette, intro card, narrator, card style) on episode 1. After that, episodes are mostly content edits, not look edits. Re-deciding the look every episode kills the cadence and burns budget.

try one of these prompts

Steal these.

faq

Common questions.

Not via URL. Paste the text directly into chat and ivy will work from that.

Your first one's on us.

Paste a URL, get a video back in three minutes. Two free credits to start.