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

# Incident Logging Control

> Turn off incident storage so Tuteliq keeps no record of your detections, and set how long incidents are retained

Tuteliq stores an incident record whenever a detection warrants attention, so you can review a moderation queue, run trend reports and produce compliance evidence.

If you already run your own moderation infrastructure, you can turn that off. Detection results still come back in full on every API response; Tuteliq simply keeps no record of them.

<Note>
  This setting controls **incident metadata** only. It does not change Tuteliq's content policy: original text, images, audio, video and documents are never stored, regardless of this setting.
</Note>

## What incident metadata is

An incident record holds the abstracted result of a detection, never the content that produced it:

| Stored                                                    | Not stored                                     |
| --------------------------------------------------------- | ---------------------------------------------- |
| Detection result and confidence                           | Original text, images, audio, video, documents |
| Risk category and severity                                | Personal information from the content          |
| Detected tactics and patterns                             | Conversation context or surrounding messages   |
| Your own identifiers (external ID, customer ID, platform) | Anything beyond the identifiers you supply     |

Summaries are encrypted. If you have registered a public key for bring-your-own-key encryption, they are hybrid envelopes that Tuteliq cannot decrypt.

## Turning it off

The setting is `incident_moderation_enabled` on your account, and it applies to every detection call made with that account's credentials. It defaults to `true`.

<Steps>
  <Step title="Open account settings">
    Go to **Settings > Incident Logging** in the [dashboard](https://tuteliq.ai/dashboard).
  </Step>

  <Step title="Switch incident logging off">
    The change applies to new API calls within one minute. Existing incidents are not affected and remain available until they expire under your retention period.
  </Step>
</Steps>

There is no per-request override. This is an account-level setting, and detection request payloads are unchanged by it.

## What changes when it is off

Nothing is written, rather than written and then deleted. No record is created at any point.

|                                      | Logging on      | Logging off        |
| ------------------------------------ | --------------- | ------------------ |
| Detection result in the API response | Yes             | **Yes, identical** |
| Incident record stored               | Yes             | No                 |
| Grooming assessment record stored    | Yes             | No                 |
| In-app incident notification         | Yes             | No                 |
| Incident alert email                 | Yes, if enabled | No                 |
| Dashboard, incident queries, exports | Yes             | No new incidents   |
| Original content stored              | Never           | Never              |

Detection behaviour, accuracy and scoring are unaffected. The only difference is whether a record survives the response.

<Warning>
  With logging off, the API response is the only copy. If you need a record, capture it from the response and store it in your own systems. Tuteliq cannot recover an incident that was never written.
</Warning>

## Retention

While incident logging is on, incidents are retained for **90 days** by default.

You can set your own period with `incident_retention_days`, anywhere from **7 days to 2 years**. When the period expires, records are permanently deleted.

```json theme={"dark"}
{
  "incident_moderation_enabled": true,
  "incident_retention_days": 180
}
```

Retention applies to incident and grooming assessment records. Records Tuteliq holds under a statutory obligation, such as audit logs, consent records and breach logs, are set by that obligation and are not configurable.

## Choosing a setting

**Leave it on** if you want a moderation queue, historical trend analysis, or evidence for a DSA, KOSA or COPPA review. The [audit receipt](/gdpr) for a specific request is the artefact to cite when a regulator asks for proof that a particular assessment happened.

**Turn it off** if you have your own logging infrastructure and would rather no vendor hold a second copy. A shorter retention period is the middle ground: set it to 7 or 30 days and keep incidents just long enough to work a queue.

## Data subject requests

While incident logging is on, Tuteliq processes incident metadata as a processor on your behalf, and you remain the controller. See [GDPR compliance](/gdpr) for the export, rectification and erasure endpoints.

While it is off, there is no incident data to export or erase, because none was created.
