Want to access all premium content?
Become a paying subscriber to get access to all courses, guides like this, prompts and skills library, and $500+ worth of AI tools.
Three Grok Bot templates can help turn a long recording into a finished clip pack. Clip Bot handles the video, Stills & Clips Desk pulls images, and Copy Humanizer works on the post text.
This guide sets up the team and takes one recording through the process. It covers production; finding clients and setting prices are still your job. You'll need an eligible Grok Bot plan, the desktop app, and a recording you're allowed to clip.
Step 1: Choose the roles. Give each bot one part of the pack.
Step 2: Check the templates. Read what each bot does before importing it.
Step 3: Create the group. Keep requests and handoffs in one chat.
Step 4: Set the handoffs. Explain when each bot starts and stops.
Step 5: Test a clip pack. Review the files and set usage limits.
Step 1: Choose the roles
Start with a clear deliverable: three captioned clips, one thumbnail per clip, and matching post drafts.
Divide the work like this:
Bot | Job | Output |
Clip Bot | Finds moments and cuts approved clips | Captioned clips with source timestamps |
Stills & Clips Desk | Pulls and sizes stills for each destination | Images, captions, and alt text |
Copy Humanizer | Rewrites post text in your voice | Drafts with a list of changes |
Before starting, check that:
Your plan includes Grok Bot. Eligible plans include Cursor Pro, SuperGrok, and others.
The macOS or Windows desktop app is installed.
You own the recording or have permission to clip it.
Find bot templates at x.ai/bot/marketplace. The Marketplace inside the app lists plugins.
Step 2: Check the templates
A template provides a starting setup. Anyone can publish one, so read it before adding it.
All your bots use the same computer, files, browser sessions, and logins. Separate bot screens don't restrict access to those resources.
Check three things on each template page:
The bot's job and where it stops.
Any files it keeps.
Its plugins and routines.
These templates have different boundaries. Clip Bot shouldn't invent quotes, speakers, or timestamps. Stills & Clips Desk doesn't choose moments, recut videos, or create images that weren't filmed. Copy Humanizer doesn't publish posts.
Select Import Bot on the template page. The documentation calls this Add to Grok Bot. Review the template in the app, then select Add Bot. Repeat for all three.
Each import creates a separate bot. Plugins reinstall, but custom code and non-standard plugins don't transfer.
Check each bot after importing:
List your skills and routines. Identify which ones came with your template and anything missing from the template description.
A reported import bug can leave skills missing, and the changelog through version 0.66.0 doesn't mention a fix. Compare the bot's answer with its template page before continuing.
Step 3: Create the group
A group chat keeps the team's requests and handoffs visible.
In New chat, select Clip Bot, Stills & Clips Desk, and Copy Humanizer. Groups support two to six bots. Name this one Clip Pack.
On iPhone, tap + and choose New Group Chat.
You can write a general message and let the bots decide who responds. For a specific job, type @ and select its owner. Use @everyone for instructions that apply to the whole team.
Step 4: Set the handoffs
Answer each bot's setup questions in its individual chat.
Clip Bot: timezone, recording type, audience, platform, clip length, aspect ratio, and spoken language. Its defaults are vertical 9:16 clips lasting 20 to 90 seconds.
Stills & Clips Desk: assets you need, destinations, sizes, visual rules, excluded content, and caption style.
Copy Humanizer: examples of your writing for its voice profile, which it reads before each job.
Then paste this into the group:
This group turns one recording into a clip pack.
Clip Bot: transcribe the recording and suggest five moments with timestamps. Wait for my selection, then cut only the approved moments.
Stills & Clips Desk: pull a thumbnail from each approved clip for the destinations I specify. Include a caption and alt text.
Copy Humanizer: draft post text from each clip's transcript in my voice. List your changes.
Don't publish or send anything. Save files on the computer and include their paths in handoff messages.
Messages between bots are text only. File paths let the next bot find the clips and images.
Step 5: Test a clip pack
Upload your recording or paste its link. Mention Clip Bot and send:
Here is my recording: [LINK OR FILE]. Transcribe it and suggest five moments for [PLATFORM], with timestamps. Wait for my selection.
Choose three moments, then review the outputs:
Clip Bot: clips with burned-in captions, separate subtitle files, and source timestamps in filenames.
Stills & Clips Desk: correctly sized thumbnails with captions and alt text.
Copy Humanizer: post drafts with numbered changes and reasons.
Check quotes against the transcript and captions against the footage. Publish the finished pack yourself.
Next, open Settings, then Usage & Billing. Your plan includes weekly usage; extra usage is billed by token cost. On an individual plan, set the monthly on-demand limit to None or Fixed.
Keep routines off until the team completes a clean test. You can pause or delete them later.
Use cases
Once the test works, try these four jobs.
Use case #1: Deliver a client pack
Agree on the number of clips, platforms, and revision rounds before quoting a price.
Use [CLIENT RECORDING], which I have permission to clip.
Clip Bot: suggest five moments with timestamps and wait for my selection. Then prepare the approved clips with captions.
Stills & Clips Desk: create a thumbnail for each.
Copy Humanizer: draft posts using the client's writing samples.
Save all the files in a folder named for the client and post its path. Don't send it to the client.
Review the quotes, footage, images, and drafts before delivery.
Use case #2: Clip a podcast
Find moments that make sense without the full episode.
Use [EPISODE LINK]. Clip Bot: suggest five standalone teaser moments with timestamps and wait for my selection.
For the approved moments, prepare captioned clips, thumbnails, and post drafts for [PLATFORM].
Watch each clip to check that the cut preserves the speaker's meaning.
Use case #3: Package a product demo
Use the same recording for a launch clip, documentation image, and release note.
Use [DEMO RECORDING].
Clip Bot: cut a 30-second launch clip showing the main feature.
Stills & Clips Desk: pull a clear still of that feature for the docs and add alt text.
Copy Humanizer: draft a three-sentence release note in my voice.
Compare the still and release note with the current product before publishing.
Use case #4: Draft LinkedIn posts
Turn selected moments into posts without changing what the speaker meant.
Use [TALK RECORDING]. Clip Bot: suggest three moments with clear takeaways and timestamps.
After I choose, Copy Humanizer should draft one LinkedIn post per moment from the transcript and list its changes.
Read each draft alongside the relevant transcript passage.
Confirmed limitations
Grok Bot requires an eligible paid plan. Monitor Usage & Billing while testing, especially if on-demand usage is enabled.
Templates don't transfer custom code or non-standard plugins. Imported skills may also be missing, so check each bot after setup.
Bots use the same computer and its access. Don't paste passwords or verification codes into chat.
Bot-to-bot messages are text only; files need to be saved and referenced by path. Stills & Clips Desk won't choose moments, so approve those before the image work begins.
This setup handles production, but doesn't guarantee income. Get written permission for client recordings and check Grok Bot's terms and template details before selling the work.
Closing
Start with one recording you own. Read the templates, add the three bots, and take a small clip pack through every handoff.
Fix missing files, awkward cuts, or inaccurate drafts before adding a routine. Once the process works, you can offer clients a clearly defined pack that you've already tested.
See you in the next one :)