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

# Authentication and keys

> Personal nx_ keys, key lifecycle, which operations need a key, and rate limits

<Note>
  This is not vaporware documentation. The API and MCP server run in production
  today, serving the chat analyst every answer it gives. Public access opens in
  beta after launch: key minting and credit packs open with the beta. This page
  documents the live contract. [Launch status](/token/status) tracks what is on
  today.
</Note>

You authenticate every paid call with a personal key in the `Authorization`
header. That is the entire auth model: no OAuth, no session tokens, no
signature schemes.

```bash theme={null}
curl -H 'Authorization: Bearer nx_your_key' \
  'https://noxint.ai/api/v1/read?id=crypto:bitcoin'
```

A personal key starts with `nx_`. Pass it as a bearer token. A key never
belongs in a query string.

## Key lifecycle

You mint a personal key in the console at
[/dashboard](https://noxint.ai/dashboard) once the API beta opens. The full
key is shown once, at creation. Noxint stores only a SHA-256 hash, so it
cannot show you the key again or recover a lost one. If a key is exposed,
revoke it in the console and mint a new one. Revocation takes effect
immediately, not on the next cache flush.

You can hold up to 5 active keys per account. Minting a sixth is rejected
until you revoke one. Give each tool or agent its own key, so revoking one
never disrupts the others.

## Which operations need a key

| Access      | Operations                             | Cost         |
| ----------- | -------------------------------------- | ------------ |
| Keyless     | root, `docs`, `schema`                 | Free         |
| Keyed, free | `resolve`, `coverage`, `usage`         | Free         |
| Keyed, paid | `read`, all domain and list operations | Data credits |

The keyless operations teach the contract, so an agent with nothing can still
learn everything. The free keyed operations resolve IDs, list coverage, and
report your credits and rate limit. Only delivered market data costs data
credits. See [Data credits](/api/billing) for how charges are counted.

## Rate limits

Each personal key gets 60 rate units per minute. Most calls spend one unit.
Batch and rank requests spend units equal to their maximum reserved work: a
`read` with 10 IDs spends 10 units, a `rank` with `limit=50` spends 50. Rate
units are a separate meter from data credits, and a rejected call still spends
a rate unit. Check `usage` for your current remaining units and reset time.

<Note>
  Personal keys are for your own tooling. Reselling, redistributing,
  multi-tenant serving, and serving third parties from your key are prohibited.
</Note>

Every authenticated response is `Cache-Control: no-store`.