The Marketing Vault

Product details
YEBW
YEBWPLUS

Verified purchaser

Deals bought: 279Member since: Oct 2020
5 stars
5 stars
Nov 3, 2026

It leaves a blank before it tells a lie.

UPDATE: I am updating my review from a 4 to a 5 due to the founder’s responsiveness/kindness, receptivity to feedback, and rapid speed of implementation.

If you've ever opened an AI draft and found a testimonial you never collected, you know the small sinking feeling that comes with it. Either you catch it, or your reader does.

I bought The Marketing Vault to find out which kind of tool this is. So I tested it on my own business, against copy I'd already written by hand. My brand has strict voice rules and an honest proof situation (one sale, zero testimonials, and I intend to keep it honest), plus two very different offers: a $47 self-serve identity reading and a monthly done-for-you service for small-business owners. If a tool was going to invent something, this is where it would.

Mostly, it didn't.

Where it earned my trust:

- It would rather leave a blank than fill it with fiction. Missing proof comes back as "[NEEDS REAL CLIENT PROOF – nothing on file]", and missing policies as "[CONFIRM]". One run caught a number that wasn't in my facts, removed it, told me, and kept the original underneath in case I wanted it back. I've never seen another copy tool show its own correction like that.
- The workflows check their own work. In "Strengthen My Offer," step 1 suggested a final-sale policy and a printed add-on for a $47 digital product. Step 2 caught both, cited the EU's 14-day digital withdrawal right, and flagged a Meta ad-policy risk in my quiz hook. Step 3 noticed the earlier steps had described my deliverables beyond what's on record, and pulled them back. Three passes, each more honest than the last.
- It stays true to the facts you give it. Once my Business Facts were current, every price, deliverable and boundary came back right across roughly a dozen outputs.
- The sales page was the standout. It wrote a full long-form page for my monthly service, and its headline landed almost word for word on the one I'd already locked after weeks of work. It never saw mine. That's either very good writing or very good listening. Probably both.
- The allowance is generous where it counts. Browser runs are unlimited. Over MCP (Claude, ChatGPT and others) you get 250 generations a month, and I measured that listing tools and updating your business facts don't count against them.

Suggestions for the team (offered because I want this tool to win - UPDATE: these were all implemented within hours of posting the review)):

- Catch the two quiet slips. In about 11,000 words of copy, I found two. A nurture email said "many people I work with describe…", which turned my own founder story into reports from clients I don't have yet. A sales page promised to save "the hour" a follow-up would take, a figure nobody gave it. The next step rewrote the first one without ever naming what it was. A check that flags invented people the way it already flags invented numbers would close the gap. Until then, read every line. (You should anyway.)
- Let the voice come through. For plain, practical copy, the writing is clean and ready to use. For intimate, voice-led copy it's accurate but could belong to anyone, and each critique pass sands it a little smoother. I'd love a setting that protects the brand voice through the later passes. For now I use it for structure and honesty, and I write the heart myself.
- Tell me when my context has gone stale. My Business Context was five weeks old, and nothing nudged me. A simple "has anything changed?" before a run would protect everything downstream.
- Give the connector and the web app one shared memory. The MCP connector can save an offer ladder and a founder story, and the copy runs use both: phrases that existed only in those fields showed up in my outputs. But neither field appears on the web Facts form, and "Regenerate ICP Document" ignores both. My regenerated profile dropped two of my offers and my founder story until I copied the same text into Additional Context.
- Finish the map you've already drawn. The Growth Plan is a smart idea, and I want to use it. Right now both "Take the Growth Scorecard" controls return to the dashboard, and I couldn't find a scorecard page. "Ask your CMO" is locked for me, even though the app lists my access as Agency and the CMO panel says it's included with Agency. And the stage to-do list showed 0/5 done after I'd already completed two of those steps.

The one I'd build first: a history-aware "next three runs" panel. The app already knows what I've run. Browser runs are unlimited and MCP runs are capped, so pointing people to the right next workflow, right there in the browser, serves the user and keeps the lighter path in use.

For anyone deciding: if you care more about never publishing a made-up claim than about polish, this is your tool. The new Ask your CMO also gives great reason to consider Tier 4+ if you can swing it!

And to the team: thank you for building something that tells the truth!

Founder Team
danielpope

danielpope

Nov 3, 2026

Raven, thank you for this. You tested it the way I'd want every user to, and the writeup was precise enough that we could fix straight from it.

Everything you flagged is now addressed, fixed/changed.

- Both "Take the Growth Scorecard" buttons now open the actual Scorecard. You weren't looking in the wrong place; the links were just pointing at the dashboard. If you have existing results you'll see them, and if not the questionnaire starts right there.

- Steps 2 and 3 no longer fire on their own. "Run next" is now "Review next": it opens the step, shows your saved context, and waits for you to write your critique focus before anything generates. You only spend a run when you say so.

- The review request no longer pops up mid-workflow. It asks once, after the whole workflow saves. That fix also uncovered that a failed save could be reported as success, which is now fixed too.

- The CMO panel now says exactly who gets it: AppSumo Tier 4, Tier 5 and subscription plans. Your "Agency" label came from how early purchases were mapped, and that label doesn't add CMO access. Sorry the old message implied otherwise; that was on us.

- One shared memory. Your saved offers and founder story now live in Business Context Facts as a card that shows exactly what the AI assistant connection reads. Regenerating your customer profile passes them through word for word - the two dropped offers and your story would now survive a rebuild.

- Invented people and time. "Many people I work with," "my clients tell me," "save you the hour" - these get a "Check these claims" warning at the top when you have no client results on file and didn't write the line yourself. We flag it; we don't rewrite you. And when a later step repairs something, it says so by name: "Fixed from the previous step: removed '…'." No more silent fixes.

- History-aware next three runs. Home suggests three workflows based on what you've actually run, skips anything from the last 30 days, and points you to running in the browser - which doesn't touch your AI connector allowance. Your reasoning, built as suggested.

- Stale-context nudge and voice protection. After four weeks without a confirmed update, workflows ask "Has anything changed?" before running - one click either way. And Output settings has a per-Business "Protect my voice" switch: on, and critique passes keep your wording instead of sanding it down.

All of it went through stress testing - rapid clicks, partial inputs, hostile pasted text...and the claim checks run as permanent tests, not one-off filters.

I'd love to earn a 5 instead of just a 4. So, retest the Scorecard, the workflow steps and the new outputs - if they feel the way you wanted, consider moving it up.

If they don't, tell me what's still missing and it goes back on the list.

I'm rooting for your product too!

- Daniel

Helpful?
1
Share
Ratings