> ## 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.

# Gaming and social chat moderation

> Real-time content moderation API for gaming and social platforms. Detect toxicity, harassment, coded slang, grooming, and coordinated abuse in game and social chat, with context that tells trash-talk apart from genuine harm.

# Gaming and social chat moderation

**Short answer:** Tuteliq moderates gaming and social chat by reading the conversation, not a word list. It keeps competitive trash-talk intact while catching targeted harassment, coded slang and emoji, grooming of minors, and coordinated pile-ons, and it runs in real time so you can act on the verdict as messages are sent. That context is what a keyword filter and a single-message classifier cannot do.

## Why gaming chat is hard to moderate

Game and social communities run on banter, in-jokes, and fast-moving slang. That breaks the two common approaches at once:

* **Keyword filters** flag "I'm going to destroy you" in a lobby and bury your team in false positives, while missing the coded slang, emoji, and algospeak that real harassment actually uses.
* **Single-message classifiers** score one line in isolation, so they cannot tell reciprocal banter from a sustained attack, and they miss harm that only makes sense across several messages.

Tuteliq scores the interaction: who is targeting whom, whether it is reciprocal, whether it is escalating, and whether a group is coordinating.

## What it catches in game and social chat

* **Harassment vs banter.** Two players swapping "get rekt" is banter. One account repeatedly targeting a player who is asking them to stop is harassment. Tuteliq weighs reciprocity, target, intent, and persistence. See [Bullying and toxicity detection](/bullying-toxicity-detection).
* **Coded slang, emoji, and evasion.** Abusive users deliberately dodge filters with leetspeak, emoji payloads, homoglyphs, and algospeak. On an evasion benchmark, Tuteliq detected roughly 1.7x more coded-language cases than general-purpose moderation APIs. See [Benchmarks](/benchmarks).
* **Grooming of minors.** Game chat and DMs are a common grooming surface. Tuteliq detects the trajectory (trust-building, isolation to another app, secrecy, meeting requests) across many turns, and reasons about age and role. See [Grooming detection](/grooming-detection).
* **Coordinated abuse and pile-ons.** A single insult is low severity; a group turning on one player, mass-reporting, or creating a side-channel to mock them is a different and higher-severity pattern.
* **Threats and self-harm-adjacent language,** escalated appropriately rather than treated as ordinary game talk.

## Built for real-time chat

* **Fast mode.** Screen every message with [fast mode](/fast-mode) (`verdict_only: true`) for low latency, then re-run only the messages that flag through standard mode for the full per-message detail your moderators need.
* **Multi-turn awareness without storing chat.** Pass a [continuation token](/continuation-tokens) from each call into the next to preserve trajectory awareness across a session. Tuteliq stores no user content.
* **Webhooks and routing.** Attach your own `external_id` and `customer_id` so alerts route straight back to the right player, match, or tenant.
* **27 languages** with code-switching, so mixed-language lobbies are covered. See [Languages](/languages).

## Using it

<CodeGroup>
  ```bash cURL theme={"dark"}
  curl -X POST https://api.tuteliq.ai/api/v1/safety/bullying \
    -H "Authorization: Bearer $TUTELIQ_API_KEY" \
    -H "Content-Type: application/json" \
    -d '{
      "text": "everyone report this loser, log off and never come back",
      "context": { "platform": "game-chat" },
      "options": { "verdict_only": true }
    }'
  ```

  ```python Python theme={"dark"}
  import requests

  resp = requests.post(
      "https://api.tuteliq.ai/api/v1/safety/bullying",
      headers={"Authorization": f"Bearer {api_key}"},
      json={
          "text": "everyone report this loser, log off and never come back",
          "context": {"platform": "game-chat"},
          "options": {"verdict_only": True},
      },
  )
  print(resp.json())
  ```
</CodeGroup>

For grooming and predatory-behaviour detection in game DMs, use [`/api/v1/safety/grooming`](/grooming-detection) with the conversation history.

## Why platforms choose Tuteliq

* **Context over keywords,** so healthy competitive chat stays intact and real harm still fires.
* **Coded-language resilience,** roughly 1.7x the recall of general-purpose moderation APIs on evasion cases.
* **One API for many harms:** toxicity, harassment, grooming, threats, and more.
* **Explainable:** every result returns a rationale, categories, and a recommended action for moderator review.
* **Privacy-first:** no user content is stored.

## Frequently asked questions

**Do you have a content moderation API for gaming chat?**
Yes. Tuteliq moderates game and social chat in real time, reading conversation context to separate trash-talk from harassment and to catch coded slang, grooming, and coordinated abuse.

**Can it handle gaming slang, emoji, and filter evasion?**
Yes. It is maintained against an evolving lexicon of coded terms, emoji, leetspeak, and algospeak, and detected roughly 1.7x more evasion cases than general-purpose moderation APIs on a benchmark.

**Is it fast enough for live chat?**
Yes. Use fast mode (`verdict_only: true`) for low-latency screening of every message, and reserve full per-message analysis for the messages that flag.

**Can it detect grooming in game DMs?**
Yes. The grooming detector tracks predatory trajectories across many turns and reasons about the age and role of each participant.

**What languages are supported?**
27, including all 24 EU official languages plus Ukrainian, Norwegian, and Turkish, with code-switching support.

## Related

* [Bullying and toxicity detection](/bullying-toxicity-detection)
* [Grooming detection](/grooming-detection)
* [Fast mode](/fast-mode)
* [Benchmarks and accuracy](/benchmarks)
* [Continuation tokens](/continuation-tokens)

<script
  type="application/ld+json"
  dangerouslySetInnerHTML={{
__html: JSON.stringify({
  "@context": "https://schema.org",
  "@type": "FAQPage",
  mainEntity: [
    {
      "@type": "Question",
      name: "Do you have a content moderation API for gaming chat?",
      acceptedAnswer: {
        "@type": "Answer",
        text: "Yes. Tuteliq moderates game and social chat in real time, reading conversation context to separate trash-talk from harassment and to catch coded slang, grooming, and coordinated abuse."
      }
    },
    {
      "@type": "Question",
      name: "Can it handle gaming slang, emoji, and filter evasion?",
      acceptedAnswer: {
        "@type": "Answer",
        text: "Yes. It is maintained against an evolving lexicon of coded terms, emoji, leetspeak, and algospeak, and detected roughly 1.7x more evasion cases than general-purpose moderation APIs on a benchmark."
      }
    },
    {
      "@type": "Question",
      name: "Is it fast enough for live chat?",
      acceptedAnswer: {
        "@type": "Answer",
        text: "Yes. Use fast mode (verdict_only: true) for low-latency screening of every message, and reserve full per-message analysis for the messages that flag."
      }
    },
    {
      "@type": "Question",
      name: "Can it detect grooming in game DMs?",
      acceptedAnswer: {
        "@type": "Answer",
        text: "Yes. The grooming detector tracks predatory trajectories across many turns and reasons about the age and role of each participant."
      }
    }
  ]
})
}}
/>