Most guides stop after the render. This one covers the whole path: how to make motionmaxxing reels from a written prompt all the way to a scheduled, measurable post on TikTok, Instagram Reels, and YouTube Shorts. The generation half is well documented elsewhere. The distribution half is where most people lose a week of work, so that is where this article spends its time.
prompt -> skill output -> clip selection -> editor pass
-> media:upload -> posts:create --scheduledAt
-> TikTok / Reels / Shorts -> analytics:posts -> next prompt
What Motionmaxxing Reels Are and Why Distribution Is the Missing Step
Motionmaxxing is a way of making a plain talking head clip move: animated captions, animated logos, face tracking, b-roll cards, counters, GIFs cut into the frame, and short punchy cuts. The output is a vertical short built to hold attention in the first second. You can watch the format in action on the launch video.
The part people skip is what happens after the file lands in your downloads folder. A finished reel that lives on one platform is a fraction of the work it took to make. Three platforms behave differently on purpose:
- TikTok wants consent flags and a privacy level set before it will accept an upload.
- Instagram Reels needs a Business or Creator account, and captions cannot carry clickable links.
- YouTube turns a vertical short into a Short automatically, but custom thumbnails only apply to long form.
Creator OS handles those rules for you: you write one post, pick platforms, and it applies each network’s requirements. The generation half is covered in our Motionmaxxing agent skill walkthrough. This post picks up right after.
Before You Start: Workspace and Account Setup in Creator OS
Install the CLI and check that your key works:
npx @creatoros/cli@latest init
creatoros auth:check
creatoros accounts:list
Then connect the accounts you plan to post to. One command per platform:
creatoros accounts:connect instagram
creatoros accounts:connect tiktok
creatoros accounts:connect youtube
creatoros accounts:health
Everything you do is scoped to a workspace. An API key or CLI session pinned to one workspace cannot post to another, which matters if you run client accounts later. If you would rather work from a chat app, the hosted MCP server at https://mcp.creatoros.ca/mcp connects Claude, ChatGPT, Claude Code, Cursor, Codex, Windsurf, and VS Code without installing anything. Setup steps are at https://www.creatoros.ca/docs/mcp. API keys, the MCP server, the CLI, and the agent skills are included in every plan.
You also need the open source editing skills. They live at github.com/kevinbadi/open-edits and render locally with Python and ffmpeg. Nothing leaves your machine.
Step 1: Write the Motionmaxxing Prompt and Pick Your Reel Format
Start with raw footage: one talking head clip, phone or camera, unedited. Then decide which skill shape fits. The repo ships several, including split-animated-talking-head, fireship-style-short, and split-animated-short.
Write the prompt as a paragraph of constraints, not a wish. Here is a worked example you can adapt:
Take clip.mov (vertical 1080x1920, 62 seconds).
Style: fast, high contrast, one idea per cut.
Captions: bold white, 3 to 5 words per card, centered, pop on
at word level. No caption longer than 0.9 seconds.
Brand: drop logo.png into the top left at 40 px, opacity
0.85, fade in at 0.2 s.
Face: keep my face in frame when I talk; reframe to center
when I turn to the whiteboard.
B-roll: three cards at 0:14, 0:29, 0:44. Each card shows a
short label and holds for 1.2 seconds.
Counters: one counter animation on the stat I say at 0:36.
Music: keep the room audio under the voice, duck it by 12 dB
when I speak.
Length: target 38 to 45 seconds. Trim filler words.
End card: last 2 seconds, logo centered, handle under it.
Two things make this prompt work. First, every instruction is measurable: seconds, pixels, word counts. Second, it names the output length, which forces the skill to cut instead of padding. If your prompt has no number in it, expect a render you have to redo.
Step 2: Generate and Select Clips From the Agent Skill Output
Run the skill inside Claude Code. It writes the composition, renders locally, and hands you files. Expect a batch, not one file: alternate caption styles, a couple of different cut rhythms, maybe two b-roll placements. That is deliberate. Picking one good take from four is faster than arguing with a single render.
Choose by watching with sound off first. If the reel reads without audio, the captions are doing their job. Then watch once with sound to confirm the audio ducking and music balance.
When you have your keeper, upload it once:
creatoros media:upload reel-final.mp4
You get back a med_ id. That single id is reusable on every platform, so you upload the file once and attach it three times. Before you do anything else, check the caption length for each network:
creatoros validate:post-length --text "your caption here"
Step 3: Caption, Reframe, and Tighten the Reel in the Editor
If the render is close but not right, fix it in the editor pass rather than regenerating from scratch. The skill set includes reverse-engineer, which is useful when you want to match a reference edit you like, and the caption skills from our open source captions post cover the subtitle pass in more detail.
Practical checks before the file leaves your machine:
- Vertical crop. Anything horizontal gets letterboxed and looks amateur on Reels and Shorts.
- Caption safe zones. Keep text away from the bottom 15 percent and the right edge, where app UI sits.
- First frame. It should show a face or a bold claim, not a logo.
- Audio. Voice louder than music, and no clipping at the cut points.
- Brand consistency. Same logo position and caption font every time. This is what makes a feed look like a channel instead of a folder.
If you have never touched a timeline, that is fine. The skills render the edit for you. Your job is judgment: is the hook clear, is the pacing right, does the last frame give someone a reason to follow. You can see these decisions narrated on the KevBuildsApps YouTube channel.
Step 4: Schedule the Reel to TikTok, Reels, and Shorts in One Dashboard
This is the step almost nobody writes about, and it is the reason this post exists. One command, one file, three platforms, one scheduled time:
creatoros posts:create \
--text "Three edits that make a talking head watchable. Which one do you use? #motionmaxxing" \
--platforms tiktok,instagram,youtube \
--media med_abc123 \
--title "Motionmaxxing a talking head in 40 seconds" \
--scheduledAt 2026-03-04T18:30:00Z \
--timezone America/Toronto \
--draft
Drop --draft when you are confident and it goes out at that time. Notice what you did not have to do. No separate TikTok consent dialog, no separate Instagram business account check, no separate YouTube title field. Creator OS applies each platform’s rules per network. Short form vertical video becomes a Short on YouTube without extra work. If one network in your list is not connected, it comes back in missing_platforms and the rest still publish, so a broken TikTok connection does not silently kill your Reels post.
For TikTok specifically, you can also send the post to the Creator Inbox and finish it on your phone with trending audio:
creatoros posts:create \
--text "Same reel, finished in the app with trending sound" \
--platforms tiktok \
--media med_abc123 \
--draft
That flow matters because the API cannot attach trending audio. Bake sound into the file, or draft it and add the track by hand. Either way, one dashboard holds the whole calendar. To check what is coming:
creatoros posts:list --status scheduled
creatoros posts:list --status draft
creatoros analytics:best-time --platform instagram
Best time to post gives you a real time to paste into --scheduledAt instead of guessing. If you run more than one brand, the Agency plans support 5 sets of socials for $49.99/month and 10 sets for $99/month. A single creator fits inside the Creator plan: 11 platforms plus Skool, one account on each. If Reels and Shorts are your whole strategy and you want a no-code route to cross posting, our one dashboard scheduling guide shows the same flow from the web app.
Step 5: Read Analytics and Iterate on Your Next Motionmaxxing Reel
Two days after publishing, pull the numbers per post:
creatoros analytics:posts --platform instagram --sortBy engagement
creatoros analytics:daily --platform tiktok
creatoros analytics:best-time --platform tiktok
Then compare the caption that came with the prompt against the analytics: a post can be perfectly edited and still miss if the copy does not land. The Instagram trial Reels post covers the same problem, testing a reel on non-followers before it hits your main feed, which is a useful pattern when you are iterating on hook styles.
What to actually change on the next pass, based on what you see:
- Low watch time in the first two seconds. Your hook frame failed. Move the strongest claim to 0:00.
- Good watch time, low follows. The end card is weak, or you never told people what you do.
- High saves, low shares. Your reel is useful but not quotable. Add a one line claim people want to send to a friend.
- Comments asking the same question. That is a keyword automation waiting to happen, and comment to DM funnels are supported on Instagram and Facebook.
Keep a running prompt file. Every iteration adds one constraint that fixed a specific weakness. After five reels your prompt is a house style, and generation stops being a gamble.
Common Mistakes That Stall Motionmaxxing Reels After Generation
The failures here are boring and fixable.
- Rendering a square or horizontal file. Regenerate vertically before upload. Fixing this after the fact means a crop that cuts captions.
- Uploading the same file three times. Upload once with
creatoros media:uploadand reuse themed_id on every platform. - Scheduling to a platform that is not connected. Run
creatoros accounts:healthweekly. A silent disconnect shows up as a failed post, not a warning. - Expecting trending audio from the API. It cannot attach it. Draft the post and finish the sound in the TikTok app, or bake the audio into the render.
- Publishing a draft as a test. Use
--draftwhile you tune the caption, then schedule. Drafts also let you eyeball the per platform preview fields before anything goes live. - Ignoring the first frame. Most viewers decide in one second. A logo intro loses them before the content starts.
Get Started With Creator OS
Install the CLI, connect three accounts, and publish your next reel to all of them from one command:
npx @creatoros/cli@latest init
creatoros accounts:connect tiktok
creatoros media:upload reel-final.mp4
creatoros posts:create --text "your caption" --platforms tiktok,instagram,youtube --media med_yourid --scheduledAt 2026-03-04T18:30:00Z --timezone America/Toronto
The Creator plan is $14.99/month or $59.99/year, with 11 platforms plus Skool, one account on each, and it includes the CLI, the hosted MCP server, API keys, and all 14 agent skills. Create an account at https://www.creatoros.ca/sign-up. For everything not covered here, including the REST API base URL, post options per network, and the full command reference, start at https://www.creatoros.ca/docs. If you want the video walkthrough of the generation half, watch the KevBuildsApps channel.
Get started for $0
Free trial for new subscribers. Cancel anytime.