FAQ

Do I need to know how to code?

No. You talk to the toolkit in plain English. The only mildly technical part is the one-time install, and the install guides walk you through every click and command.

Cowork or the desktop app, which should I use?

Use Cowork for research, strategy, copywriting, and the whole account-operation layer including the weekly routine, audits, diagnostics and scaling proposals (no setup, no terminal). Use the Claude Code desktop app when you need to generate ad images or video, or build and deploy full landing pages. Many people use both. Full breakdown on Choose your setup.

Why can’t I generate images or video in Cowork?

Cowork runs in a secure sandbox that currently blocks the outside services those tools call to render. The image and video skills can still plan and write prompts in Cowork, but to produce finished files you need the desktop app. We expect this gap to narrow over time.

Is my brand data private?

Yes. Your brand files live in your own workspace folder on your computer. They aren’t shared with other users, and toolkit updates never touch them.

Will updating the toolkit erase my brand setup?

No. Updates only refresh the skills, rules, and templates. Your offer, voice, personas, and everything else you set up stay exactly as they are. See Updating.

Can I use it for more than one client?

Yes. Set up each brand separately and the toolkit keeps them fully isolated, so one client’s voice or data never leaks into another’s work. Add a new one any time with the onboard-client skill.

Does it cost anything to run?

You use it through your own Claude account. The only extra costs are your own third-party API keys for image and video generation (Gemini and Creatify), and only if you use those skills. Everything in the research, copy, strategy, and review group needs no extra keys. See Connect your API keys.

Will it change anything in my ad account?

No, not without you approving that specific change. Every skill that touches an ad account is read-only. Proposed changes ship as a list you approve one diff at a time, and there is deliberately no approve-all. See meta-scaling-copilot.

Does it need access to my ad accounts?

Only for the account skills (the weekly routine, the audits, the diagnostics, scaling). Read-only access is enough. Everything in research, copy, offer, angle and creative works with no account access at all. If a surface isn’t connected, you get a specific setup message naming what’s missing and what the run degrades to, rather than a vague error.

What does NO-DATA mean in a report?

It means that check could not be evaluated, usually because a data source wasn’t reachable. It blocks exactly like a failure does, so a run that couldn’t see your data never comes back looking clean. It’s the single most important thing to understand about reading output: gates and verdicts.

My business isn’t ecommerce. Will the numbers make sense?

Yes, once you declare your money path. One line in your brand guide tells every skill whether you’re an ecom store, a lead-to-close business, or a booking business, and the arithmetic follows it. A consultancy with a 45-day sales cycle gets a two-tier banked-plus-leading headline instead of ecommerce math with the words changed. See conversion shapes.

Will it make things up?

No. The toolkit is built to flag gaps instead of inventing facts. If it can’t verify a stat, a testimonial, or a claim, it tells you rather than filling it with something that sounds good but isn’t true. You should still sanity-check specifics before anything goes live.

The output sounds generic. How do I fix it?

That’s almost always thin brand files. Re-run onboarding and give richer answers, especially 3 to 5 real samples of your brand’s actual writing. Voice samples are the single biggest lever on how much the output sounds like you.

How do I get help?

Start with Troubleshooting for the common issues. For anything else, ask in your team channel.