> ## Documentation Index
> Fetch the complete documentation index at: https://docs.limitguard.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# September 5, 2026

> Documentation corrections — pricing, tier ladder, and MCP tool names

## Documentation Corrections

A sweep of the documentation site against the live API. No API behavior changed — these entries correct places where the docs had drifted.

<Warning>
  If you copied code samples from these docs before this date, check your hard-coded x402 amounts. The previously documented `50000` (\$0.05) amount for `/v1/entity/check` was wrong and will now be rejected as insufficient payment.
</Warning>

### Pricing corrected

Per-call prices across all pages now match [Pricing](/pricing). The most visible change is `/v1/entity/check`, previously documented at $0.01 / $0.05 / \$0.15 across the cached, fresh, and enhanced quality tiers:

| Endpoint                | Cached | Fresh  | Enhanced |
| ----------------------- | ------ | ------ | -------- |
| `POST /v1/entity/check` | \$0.10 | \$0.85 | \$1.50   |
| `POST /v1/risk/score`   | \$0.10 | \$0.65 | \$1.00   |
| `POST /v1/kyb/check`    | \$0.25 | \$1.50 | \$3.00   |

Every x402 `amount` field, `X-Payment-Amount` header sample, and executable tier-price table in the guides was updated to the 6-decimal equivalents.

### `enterprise` tier removed

The API key tier ladder is `free` / `sandbox`, `indie`, `starter`, `growth`, `pro`. There is no `enterprise` tier — it has been removed from the rate-limit and error-handling tables. Pro's monthly quota is 250,000 calls, not 50,000.

Only `free` and `sandbox` keys can be provisioned directly through `POST /v1/keys/create`. All other tiers are reached by paying via x402 against `GET /v1/keys/upgrade/{tier}`. See [Authentication](/authentication#api-key-authentication).

### MCP tool names

<Note>
  Superseded by [September 7](/changelog/2026-09-07): two of the five names below are
  wrong. The MCP server answers to `get_trust_score` and `get_risk_score`, not
  `trust_score` and `risk_score`. The paragraph is left as it was published, because a
  changelog entry records what was said on a date.
</Note>

The manifest at `/.well-known/mcp.json` declares `check_entity`. Documentation examples that called it `check_entity_trust` have been corrected. The five tools are `check_entity`, `check_agent`, `trust_score`, `verify_wallet`, and `risk_score`.

### Discovery response shape

The `/.well-known/x402.json` examples now match the live response. Endpoint entries carry `price_usdc`, `price_raw`, and `category`. The listing does not include a `quality_tiers` object — per-tier prices come from [Pricing](/pricing#tiered-pricing) or from the `amount` in the 402 response.