# Every claim checked before you see it

> AI writes fast and sometimes invents things. Brandiz.io doesn't let it: every number, price and claim in a post must match a fact you confirmed, with the page it came from. A post that fails is fixed or held back.

Web page: https://www.brandiz.io/features/truth-checks · Updated 2026-10-21

## How the check works

1. **Facts from your pages**: When the team reads your website it keeps each fact with the page it came from: prices, numbers and what you offer.
2. **Every word is read**: Before a post reaches you, code reads every word. A number without a source is flagged, and a banned claim stops the post.
3. **You decide**: A flagged number shows you why. Fix the words, or confirm the fact with one tap so the team can use it again.

## What it catches

### Numbers with no source

“3x faster”, “500 customers”: a number that reads like a claim needs a fact you confirmed.

### Prices as your store shows them

A product poster quotes the price exactly as your store shows it, sale and sold-out included.

### Words you never want said

Add phrases your business never uses, and any post that uses one can't be approved until it changes.

### Numbers in Urdu too

A number written in Urdu digits is read like any other, so the check can't be passed by changing script.

## What we never write

- Avoid: Reviews or testimonials.
- Avoid: Star ratings, prizes or rankings you don't have.
- Avoid: Promises of results, like more followers or more sales.

## It reads your own words too

The same check reads the captions you write yourself before they can be approved. It protects your name: a post with your logo on it should only say what you can stand behind.

It catches numbers and claims without a source. It can't judge taste or tone, which is why every post still waits for your yes.

## Questions

### What happens when a post fails the check?

A post with a banned claim can't be approved until the words change. A number without a source is flagged with the reason, and you either fix the words or confirm the fact with one tap.

### Does it check posts I write myself?

Yes. Everything that can be published runs the same check, including captions you write.

### Can the AI still get something wrong?

Yes. The check catches numbers and claims without a source, but it can't judge taste or tone. That's why nothing goes out without your yes.

## Read next

- [How we check](https://www.brandiz.io/how-we-check)
- [Approvals](https://www.brandiz.io/features/approvals)
- [Brand kit from your website](https://www.brandiz.io/features/brand-kit-from-website)