> For the complete documentation index, see [llms.txt](https://docs.tallyprotocol.xyz/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.tallyprotocol.xyz/concepts/integrity-grades.md).

# Integrity grades

Every token gets a grade from A to F before you can buy it, and every point taken off is shown with its reason. The grade is computed by the pure engine in `packages/core`, so the web app, the Telegram bot and agents all see the same grade.

## The scoring

Start at 100 points and subtract:

| Condition                                                      | Points | Notes                                           |
| -------------------------------------------------------------- | ------ | ----------------------------------------------- |
| Multiplier sources disagree by more than 0.1%                  | 25     | List, API and chain readings are compared.      |
| Price more than 2% away from the US price in regular hours     | 30     | Premium is measured per share, never per token. |
| Under $1,000 of onchain volume in 24 hours                     | 40     | Flagged as a **ghost market**. Not executable.  |
| Trading status unknown                                         | 10     | A missing status is never assumed open.         |
| Token paused by the issuer                                     | 50     | Not executable right now.                       |
| Token limited (for example around earnings)                    | 10     |                                                 |
| Issuer's latest reserve report older than 3 days               | 10     | Only when a dated report exists.                |
| One token is more than one share while another issuer's is one | 0      | Adds the **unit trap** badge, not a deduction.  |

| Grade | Points    |
| ----- | --------- |
| A     | 90 to 100 |
| B     | 75 to 89  |
| C     | 60 to 74  |
| D     | 40 to 59  |
| F     | below 40  |

In the app, A and B read **Liquid**, C to F read **Low Liquidity**, and a ghost reads **Not Tradable**.

## Two grades, on purpose

* The **quote grade** on the Trade page uses raw 24-hour onchain volume, so it can be computed instantly for every quote.
* The **Radar grade** uses *cleaned flow*: trades left after bot wallets and router hops are removed. The two can differ, and the Radar page says which basis it uses. We chose not to wire the cleaned figure into the quote path, so the two never silently disagree inside one screen. See [Radar and Flow](/product-modules/radar-and-flow.md).

## The integrity log

Every check writes a record (inputs, outcome of pass, deduct, flag or skipped, and a one-line summary) whether or not it costs points. A fact that is missing carries the reason, and the quote shows it as a warning. The rule we followed throughout: **never add a check or a fallback that can fail silently.** An early adapter swallowed a schema failure and fell back to stale public data; every fallback now reports a warning.

## Checks against the issuer's own claims

The Ondo multiplier has no onchain source, so every reading is bounded:

* An increase of up to 3% is accepted unless the independent price check fails. The 3% is a judgement threshold, about five times the largest step observed (+0.58%).
* A decrease, or an increase above 3%, is accepted only when a split or dividend status was seen within 48 hours, the ratio is within 0.5% of a simple ratio (2, 3, 4, 5, 8, 10, 15, 20, 25, 30, 50, 3/2, 2/3 or their inverses), and the token price divided by the new multiplier is within 2% of the US share price.
* Otherwise the token stays blocked and flagged.

Only readings that pass are stored as the new baseline, so a bad reading cannot become the reference.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.tallyprotocol.xyz/concepts/integrity-grades.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.