Runtime Text To Speech (Real-Time, Offline, Streaming TTS) is a code plugin for Unreal Engine that performs offline text-to-speech synthesis in real time, using Piper, Kokoro and ONNX Runtime. It supports 51 languages and 2800+ voices.
Runtime Text To Speech is a Unreal Engine plugin for generating spoken dialog directly inside your game, without an internet connection or a cloud service. It is built on the Piper and Kokoro voice models and the ONNX Runtime library, and it supports 51 languages, 2800+ voices and 75 voice qualities. The plugin is distributed as a code plugin for Unreal Engine 5.0 through 5.8 and targets games that need dynamic speech, such as NPC conversations, voice-driven tutorials, accessibility features and procedurally generated dialog.
Specifications
| Seller | Georgy Dev |
|---|---|
| Type | Plugin |
| Category | Dialog Systems |
| Engines | Unreal Engine 5.0–5.8 |
| Distribution | Code plugin |
| Formats | Unreal Engine |
| File types | .uasset, .uproject, .uplugin |
| Unreal target platforms | Android, Linux, Mac, Oculus, Windows, iOS |
| Licenses | Personal, Professional |
| Published | 2026-10-07 |
Key features
- Offline text-to-speech synthesis with no internet connection required
- 51 languages and 2800+ voices with 75 voice qualities
- Kokoro voice models with 151 models across 8 languages
- One-click voice model downloads and in-editor voice preview
- Runtime voice model selection and custom Piper and Kokoro model import
- Raw PCM float audio output for any playback solution
- Full Blueprint and C++ API with regular and streaming synthesis
- Synthesis cancellation and automatic voice model packaging
What the Plugin Includes
The plugin lets you download and manage voice models from the editor, preview voices in the editor, and package the chosen models with your project. You can select a voice model at runtime from a dropdown and import custom voice models in Piper and Kokoro formats.
Synthesis runs in regular and streaming modes. Streaming mode returns real-time audio chunks, and synthesis can be cancelled. The output is raw PCM float audio, so you can connect it to any audio playback solution. The plugin is also compatible with Runtime Audio Importer, which the creator uses in its own examples and demo project.
Full Blueprint and C++ APIs are available, along with access to voice metadata.
Platforms, Engine Versions and Distribution
Runtime Text To Speech supports Unreal Engine 5.0, 5.1, 5.2, 5.3, 5.4, 5.5, 5.6, 5.7 and 5.8. Its Unreal target platforms are Windows, Linux, Mac, Android (including Meta Quest), iOS and Oculus. Experimental support for Apple Vision Pro is also mentioned by the creator.
The plugin is delivered as a code plugin and uses the .uplugin, .uproject and .uasset file types. Licenses are offered as Personal or Professional. Support for Unreal Engine 4.27 is available only on request.
Use cases
- Dynamic NPC conversations
- Voice-driven tutorials and hints
- Accessibility interfaces
- Procedurally generated dialog
- Localized game dialog
- Interactive storytelling
Runtime Text To Speech (Real-Time, Offline, Streaming TTS) FAQ
Which Unreal Engine versions does Runtime Text To Speech support?
The plugin supports Unreal Engine 5.0 through 5.8. Support for Unreal Engine 4.27 is available only by contacting the creator.
Which file types are included with the plugin?
The plugin uses .uasset, .uproject and .uplugin files and is distributed as a code plugin.
Which platforms can a project using this plugin target?
Unreal target platforms are Windows, Linux, Mac, Android (including Meta Quest), iOS and Oculus.
Does the plugin need an internet connection to generate speech?
No. Synthesis is offline and runs in real time without internet connectivity.
Which license types are available?
The plugin is offered under Personal and Professional licenses.

