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

# API Docs: Voice

## **Introduction**

GGWP Voice Moderation extends GGWP’s moderation platform to voice-based player interactions.\
The service ingests audio clips through a straightforward API endpoint, processes them through a game-tuned speech-to-text pipeline, and evaluates the resulting transcripts using GGWP’s moderation models and policy engine.

Voice moderation results flow into the same GGWP platform used for text chat, reports, Discord, user reputation, dashboards, webhooks, Auto-moderation, and Player Alerts. This gives teams a consistent view of player behavior across interaction channels.\
\
GGWP Voice Moderation supports:<br>

* Unified user reputation across Voice, Text Chat, Reports, and Discord
* Detection across GGWP’s supported incident taxonomy
* Dashboard review, webhook events, and user management workflows
* Auto-moderation and Player Alerts
* Secure, compliant, and privacy-conscious handling aligned with GGWP’s broader data services

GGWP also applies lightweight audio-side processing, such as speech activity detection, triage, keyword or phrase spotting, and signal extraction, to support efficient routing and prioritization. Core moderation decisions are based on transcript analysis, policy configuration, and available user context within the GGWP platform.

## **Language Support**

GGWP's Voice Moderation service offers support for the following languages:&#x20;

* Chinese (simplified and traditional)
* English
* French
* German
* Indonesian
* Italian
* Japanese
* Korean
* Portuguese
* Russian
* Spanish
* Turkish

New languages are added regularly. Please reach out to your Account Manager to learn about upcoming languages or submit additional languages for consideration.

## **Model Support**

The categories covered by GGWP's Voice Moderation are described in the [model glossary section](/text-and-voice-model-glossary.md).

## Integration Workflow

Because our service requires only a single Voice Ingestion API call, there are a number of ways to quickly integrate regardless of a game's voice infrastructure. An example path looks like this:

1. Identify where user audio files or streams can be accessed.
   1. If audio is already being stored somewhere (e.g. AWS S3, internal data warehouse, third-party service), then no further engineering is needed.
   2. If audio is not being stored, then it can likely be accessed as audio streams directly in the "audio rooms" created by voice services like Vivox, EOS, Photon, and ODIN for each voice session. Using the service SDK, it is possible to chunk these audio streams and either feed them directly to our ingestion API or save them for offline processing. Examples:
      1. EOS describes [here](https://dev.epicgames.com/docs/game-services/real-time-communication-interface/voice/using-voice#speech-to-text) how audio streams can be sent to 3rd-party transcription services which is identical to how GGWP voice ingestion works.
      2. Vivox describes [here](https://support.unity.com/hc/en-us/articles/4418149054356-Vivox-How-to-Access-client-side-audio-buffers) how callbacks can be used to process audio streams during various stages of client / server interactions, such as storing or sending to an API service.
      3. Photon describes [here](https://doc.photonengine.com/voice/current/troubleshooting/faq#how_to_save_conversations_into_files_) how audio streams can be saved to files in a similar manner.
2. Process user audio into file chunks conforming to the requirements in Voice Ingestion API.
   1. Requires light coding for which GGWP can provide snippets and implementation help.
3. Submit audio files through the ingestion API where most convenient (either client or server-side).
4. Access user profiles and incidents shortly after through GGWP Dashboard or webhook events.