# Needs a person (`mttr_reviews`)

Updated October 24, 2026.

## What does mttr_reviews do?

Verdicts a model could not be relied on for, which wait for a person and count neither for nor against a score, newest read first. Only whoever runs this mttr may make this call.

Who may call it: Owner.

## When to use it

- You want to clear the verdicts that wait for you.

## What you ask

You might say: “What is waiting for a person on mttr?”

```
mttr_reviews({})
```

| Argument | Type | Needed | What it is |
| --- | --- | --- | --- |
| `offset` | Number | No | How many to skip, for the next part of a long list. |

## What comes back

- Each site with something waiting, and the total across every site.
- What the model said, marked as quoted.

A sample, with made-up sites and figures:

| Site | Rule | Page |
| --- | --- | --- |
| example.com | Answer the first question right under it | /guides/septic-care |
| shop.example | Replace unsupported sales claims | / |

## Over HTTP

```
curl "https://mttr.si/api/v1/reviews" \
  -H "Authorization: Bearer $MTTR_KEY"
```

## When it fails

| When | Status | What to do |
| --- | --- | --- |
| Your assistant has no key yet | 401 over MCP | Sign up with `mttr_signup`, then connect again with the key it gives. |
| The key is a person’s, not the owner’s | 403 | Nothing: only whoever runs mttr makes this call. |

## Sources

- [Model Context Protocol: tools](https://modelcontextprotocol.io/specification/2025-06-18/server/tools), Model Context Protocol: How an assistant calls a tool.

The call in full: [mttr_reviews in the API reference](https://mttr.si/docs/reference#mttr_reviews).

## Goes with

- [The rulebook](https://mttr.si/docs/tools/rules.md)
- [One site](https://mttr.si/docs/tools/site.md)