For $25 a month, I can now run all my social media by myself!
No hired video editor. No agency. No outsourced distribution.
I recently told friends this, and they immediately asked me to share the workflow. So here it is.
Most of it runs through Codex, connected to @chatcutapp for video editing. I use @bot and @Muse for distribution.
Here’s how it works.
1. Written whenever inspiration hits.
One idea can become a written post, a short video, and a longer YouTube video.
Sometimes I dictate because I’m excited about something and have far too much to say. You can use a voice-dictation tool like @typelessdotcom for this. Sometimes I’m typing notes on my phone whenever I have a few spare minutes. You might catch me writing up a storm on @Caltrain.
At that point, the material can be messy. There might be an argument without enough evidence, a reference I need to track down, or a fact that needs checking.
I ask @Codex to help research those gaps, find sources, and verify the claims. Then I review and revise the draft.
Once the written version is clear, I already have most of what I need for a video.
This stage can feed into Substack or Medium. A tool such as @Statkraft can prepare a Markdown draft as formatted text to paste into their editors.
2. Record whenever I have time.
I already have the script, so recording becomes much easier to fit into my day.
I take out my phone, load the script into a teleprompter, and record. If you need an English-language option, @BIGVU lets you read a scrolling script while recording on your phone.
3. Let Codex handle the editing through @chatcutapp.
I used to hire video editors. Now I use the ChatCut plugin inside Codex for both my short videos and my longer YouTube videos.
It handles several parts of the process:
• Speech cleanup: removing failed takes, repeated starts, and excessive pauses.
• Captions: transcribing the recording, applying a caption style, and checking the wording against the script we already worked on.
• Supporting visuals: finding relevant material from my existing files, looking for additional sources when needed, and placing the visuals alongside the right part of the explanation.
• Composition: arranging screenshots, source material, captions, and my talking-head footage so the important information stays readable.
If I’m explaining a screenshot, I want people to see the screenshot. My face can move into a smaller window. The captions need to stay clear of both my face and the information I’m showing.
If I see a layout I like, I can give Codex a screenshot as a reference for the colors, typography, and composition, then ask it to build titles or motion graphics through ChatCut.
I also save my recurring preferences in the Skills I wrote:
• How the opening title should look.
• How captions should be grouped.
• Where supporting images should go.
• Which photos and text belong in the closing.
• How to keep the whole video visually consistent.
Once those instructions are written down, I can reuse them for the next video.
4. Schedule the finished content.
I use @bot to schedule posts on X, and @Muse to schedule Instagram and Threads posts.
My core subscription setup is:
• ChatGPT Plus, which includes Codex: $20/month.
• ChatCut: $25/month.
I open-sourced my three skills: writing, short video editing, and long video editing.
The links are in the first reply. You can adapt the instructions to your own voice, visual style, and publishing platforms.
If you try them, I’d love to see what you make.