Big Idea

The argument in one line.

Bolting a social-publishing API onto an open-source AI video editor's timeline turns a personal editing tool into a full pipeline that renders and posts content from one chat prompt, closing the gap between finishing an edit and having it live.

Who This Is For

Read if. Skip if.

READ IF YOU ARE…

You're already running (or plan to run) a self-hosted or open-source AI video editing setup and want to see how automated social publishing gets bolted onto it.
You're comfortable cloning a GitHub repo, pointing a coding agent (like Cursor) at it, and wiring in an API key to get a local dev server running.
You're interested in agent pipelines that chain several tools together (FFmpeg, Remotion, an ingestion pipeline, a social API) rather than a single-purpose app.

SKIP IF…

You want a polished, no-code social scheduler you can use today — this is a developer-facing open-source project with real setup steps and two prerequisite videos.
You haven't watched HyperEdit's first two tutorials (base editor setup, Obsidian ingestion pipeline) — this video assumes both are already running.

TL;DR

The full version, fast.

HyperEdit is Kevin Badi's open-source AI video editor, built on FFmpeg, Remotion, and local coding agents, that he's iterated through three public updates. V3 adds Creator OS, a chat panel inside the editor's timeline that connects to a social media API. After pasting in an API key, the agent confirms which accounts are linked, then a single instruction like "export and upload to Instagram" renders the timeline and publishes it, with no manual export or upload step. The demo uses DeepSeek V4 Flash to scaffold and run the local dev server from a fresh git clone in under a minute. The takeaway is architectural: once editing lives in an agent-driven pipeline, distribution can be added as just another tool call rather than a separate manual workflow.

Free for members

Chat with this breakdown — free.

Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.

Create a free account →

Chapters

Where the time goes.

00:00 – 00:19

01 · Why open source

Kevin frames his use of free, open-source tools as an AI-merging advantage, not a cost-saving one.

00:19 – 00:35

02 · What HyperEdit does

Recaps HyperEdit as an AI video editor that matches CapCut's feature set (transcribing, captioning, cutting, animating).

00:35 – 01:01

03 · V1 and V2 recap

150+ people recreated the V1 codebase; V2 added an Obsidian knowledge graph for sourcing video content.

01:01 – 01:37

04 · Announcing V3: Creator OS

Introduces the V3 update: a social media AI layer that uploads edited timeline content directly to social accounts.

01:37 – 02:09

05 · Encouragement to follow along

Kevin notes he isn't a technical engineer and encourages viewers to attempt the setup themselves.

02:09 – 02:34

06 · Prerequisite tutorials

Points to two earlier videos: base editor setup (FFmpeg, Remotion) and the Obsidian ingestion pipeline.

02:34 – 03:04

07 · Setup overview and cloning HyperEdit

Describes today's process, then copies the GitHub HTTPS link, opens Cursor, and clones the repo to start the coding agent (DeepSeek V4).

03:04 – 03:27

08 · Agent launches the editor

The agent initializes the codebase and starts the local host editor server in under a minute.

03:27 – 03:51

09 · Connecting the Creator OS API key

Pastes an API key into the new Creator OS chat panel; the agent verifies six connected social accounts.

03:51 – 04:13

10 · Explaining the skill-building workflow

Kevin describes turning each manual step into a reusable 'skill' that can later be automated.

04:13 – 04:42

11 · Dropping in a demo clip

Adds a UGC reaction clip to the timeline and instructs the agent to export and upload it to Instagram only.

04:42 – 05:08

12 · Render and upload

The agent shows an interactive processing timeline, renders the clip, uploads via the API, and confirms the post.

05:08 – 05:34

13 · Verifying on Instagram and sign-off

Checks Instagram to confirm the reel is live, then pitches a 1:1 onboarding call through No Code Academy.

Atomic Insights

Lines worth screenshotting.

HyperEdit is built entirely on open-source pieces (FFmpeg, Remotion, local coding agents), and Kevin argues the value isn't that they're free, it's that AI can merge them into combinations paid tools don't offer.
Over 150 people had already recreated HyperEdit's codebase from GitHub before this V3 update shipped.
V3's core feature, Creator OS, is a chat panel added directly to the editor's timeline that can render a clip and publish it to social accounts on a single instruction.
The Creator OS agent checks the API key against connected accounts before doing anything else, surfacing exactly which platforms (in the demo: six) are actually linked.
The entire local dev environment (server, editor UI, agent chat) was scaffolded and launched from a fresh git clone in under a minute using a coding agent.
The demo intentionally uses DeepSeek V4 Flash instead of a more expensive model, on the claim that it costs about a penny a day and performs faster for this specific agent-orchestration task.
Publishing a clip to Instagram in the demo required exactly one chat instruction: "video is done, please export and upload to Instagram only."

Takeaway

Automating the last mile: editor to published post

PIPELINE DESIGN

The V3 update's real lesson isn't the editor itself, it's that distribution can be wired into an editing pipeline as just another agent tool call.

Chaining an open-source video editor to a social-publishing API turns manual export-and-upload into one chat instruction, closing the gap between 'edit is done' and 'post is live'.
Checking which accounts are actually connected before attempting an upload catches authentication problems before a render gets wasted on a platform that isn't linked.
A coding agent can scaffold and launch a full local dev environment (server, UI, chat panel) from a fresh git clone in under a minute, which lowers the setup barrier even for non-engineers.
Recording each manual step as a reusable 'skill' while building is what turns a one-off automation into a repeatable pipeline instead of a single script.
Choosing a cheap, fast model (DeepSeek V4 Flash) over an expensive one signals that this kind of tool-calling and orchestration doesn't need frontier-level reasoning to work reliably.
Open-sourcing a build and letting 150+ people fork the same codebase generates faster iteration and feedback than keeping it closed and paid.

Glossary

Terms worth knowing.

HyperEdit: Kevin Badi's open-source AI video editor, built on FFmpeg and Remotion, that can transcribe, caption, trim, and animate video the way CapCut does, but driven by AI agents.
Creator OS: A social media publishing API and app that V3 wires into HyperEdit's timeline, letting a connected account receive a rendered video directly from the editor.
Obsidian knowledge graph (V2 update): A video ingestion pipeline added in HyperEdit's second update that gives the AI model access to a large library of source videos it can pull from when building new ones.

Resources

Things they pointed at.

03:44toolHyperEdit (GitHub) ↗

01:02toolCreator OS ↗

03:52toolDeepSeek V4 Flash API

02:09channelHyperEdit setup tutorial (FFmpeg/Remotion) ↗

02:13channelHyperEdit Obsidian update tutorial ↗

05:20productKev's No Code Academy ↗

Quotables

Lines you could clip.

00:05

“The reason why I actually use open source and free tools is not because they're free. It's because AI allows you to merge all of these open source tools together.”

clean thesis statement, no setup needed→ TikTok hook↗ Tweet quote

00:20

“I built and open sourced an AI video editor that can literally do anything that CapCut can, but with the power of AI.”

concrete claim with a familiar comparison point→ IG reel cold open↗ Tweet quote

04:57

“Every single step in that process, turn it into a skill that we can turn into an automation.”

generalizable framework line for agent-building content→ newsletter pull-quote↗ Tweet quote

The Script

Word for word.

Read-along

Don't just watch it. Burn it in.

See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.

So as some of you know, a lot of the tools that I use are actually now open sourced and completely free. But the reason why I actually use open source and free tools is not because they're free. It's because AI allows you to merge all of these open source tools together.

And that creates possibilities that paid tools just simply do not provide. And on this channel, I built and open sourced an AI video editor that can literally do anything that CapCut can, but with the power of AI. From transcribing videos, captioning videos, cutting videos, trimming videos, adding animations, adding AI graphics, the list goes on.

And once I launched it into the Academy and onto GitHub, over 150 people have recreated the exact same code base. And then we came out with the V2 update, which showed you guys how to add an obsidian knowledge graph to your AI video editor. And this turned into an entire AI video engine, which I personally have been using to build for businesses and for myself internally.

It's been absolutely game -changing. And today... I have the V3 update where we are implementing Creator OS, which is our social media AI tool, which now allows us to grab the videos and content that we're creating in the video editor timeline and have them directly uploaded to social media.

And in this video, I'm going to show you exactly how it works and exactly what the V3 update is. Pretty ridiculous. So I wrote in.

Actually, before we dive in, if this is your first time interacting with something like this, I want to encourage you to actually give it your all because as daunting as it may seem this has been honestly so revolutionary for me and like all i did was put in the work I'm not a technical engineer at all.

Because what you're seeing on the screen is the two video prerequisites that I recommend watching before this one that shows you how to actually set up the first version of this and all of the tools that are associated to it like FFmpeg and Remotion that allows AI to actually edit videos and allows AI to use your computer to actually run all of this video processing.

Then I recommend watching the obsidian update because that shows you how to build a video ingestion pipeline. And this. may sound like a lot of word jumbo, but it is really powerful because then your AI model will have access to thousands of videos that it can source to make videos, right?

Today's update though, we're going to go through the exact same setup process, and then we're going to add a video to the timeline. We're going to add our creator OS API key and directly through the timeline, I'm going to show you a post that is uploaded directly onto social media without us having to. download the video or do anything else, we could just talk to our agents and have it upload the video.

So in the description down below on this video, you're going to find this hyper edit link, and that is where all of the code lives. And all you have to do is copy the HTTPS link, and then we can just clone the entire code base. So I'm going to open up cursor, just going to paste in that URL and hit enter.

Okay, lovely. So when you're set up with your cloud code agents, in this case, I'm using deep seek V4, uh, it should look something like this. And all you have to tell your agent here is to an initialize with the code base and to launch the local host editor and then everything should be set up all right so the agent's now saying that the server is up so we're gonna go half coding agent screen half video editor screen so within like a minute of forking the project we are now ready to go and so in the right hand side literally the only update is a whole creator OS agent chat So if we go into that chat, it's going to ask for your API key.

So I'm going to go and find an API key quick on creator OS. So I'm in my settings for one of my users. We're going to click copy, head back to the editor and we can just paste in that key and it's going to check if this key is actually valid.

So we're going to send that in. It's going to check and we could see that it has six social accounts connected. So this creator OS agents now is ready to go.

And this is how I'm building out the skills, right? So we're going to add a video in and we can edit the timeline and work on all this stuff. And then once we have the end product, we can say every single step in that process, turn it into a skill that we can turn into an automation.

But all I want to show you guys now is just uploading this to social media, right? So let's just go grab. a generic video.

Cool. So we have one of our AIUGC reaction clips here and we just can like drop it into the timeline and let's just assume that this is done because everything else that is in hyper edit is in the previous tutorials. So I'm assuming you have everything else set up.

So in this example, let's say we want to only upload to Instagram. I can actually like specify whatever I want in the chat as well. Same with like descriptions and captions, but let's just keep it simple.

So I'm going to say video is done. Please export and upload. to Instagram only, please.

And then the agent is going to actually show you like an interactive processing timeline, kind of like how ChatGPT will do it. So here we could see that it rendered the timeline. Now this is a shorter video, so it went pretty quick.

It then did the upload to the API and then it listed the accounts that it wanted to upload, which is Instagram, created the post and it says done. we can also ask our agent here or we can go and check instagram so let's just go ahead and check instagram cool yeah so after one refresh on the instagram we could see the reel has been actually uploaded and yeah that is literally all you have to do also if you want hands -on support setting this agent up and setting up the entire code base head over to the no code academy i will give you a 101 onboarding call and you can use that to getting this set up see you guys on the next one

The Hook

The bait, then the rug-pull.

Kevin Badi opens by explaining why he builds with open-source tools: not because they're free, but because AI lets him merge them into combinations paid software doesn't offer. That premise sets up the rest of the video, HyperEdit's third update, which adds a social-publishing layer directly into the editor's timeline.

CTA Breakdown

How they asked for the click.

VERBAL ASK

05:09product

“if you want hands-on support setting this agent up and setting up the entire code base head over to the no code academy i will give you a 101 onboarding call”

Soft pitch placed after the working demo is confirmed live, framed as optional hands-on help rather than a hard sell.

MENTIONED ON CAMERA

03:44toolHyperEdit (GitHub) ↗

01:02toolCreator OS ↗

02:09channelHyperEdit setup tutorial (FFmpeg/Remotion) ↗

02:13channelHyperEdit Obsidian update tutorial ↗

05:20productKev's No Code Academy ↗

Storyboard