Shengyue Voice guides
Local, offline narration: requirements and first setup
Local text to speech runs on your own computer. Shengyue Voice initially downloads runtime components and voice packages. Once prepared, generation, reference-voice reuse and license verification can work offline. Test compatibility, quality and waiting time on the computer you will actually use before buying. The current desktop app uses a Chinese interface.
Choose the installer for your computer
| Platform | System requirements | Current inference |
|---|---|---|
| Apple Silicon Mac | macOS 14+, Apple M-series chip | Available modes depend on the local check |
| Intel Mac | macOS 14+, Intel x64 | CPU; full native Intel acceptance is pending |
| Windows | Windows 10 2004 / 19041+; Windows 11 recommended; x64 | CPU; native installation and inference acceptance is pending |
There are no Linux or Windows ARM installers. Intel Mac and Windows builds do not include CUDA acceleration. The Mac build is not Apple-notarized, so first launch may require confirmation in System Settings → Privacy & Security. Choose the installer that matches your chip.
How much memory and storage do you need?
Light mode needs at least 8 GB RAM. Standard narration, character creation and reference-voice reuse need at least 16 GB and must pass the app’s checks. Other applications also use memory; a minimum configuration does not guarantee every task will run quickly or reliably.
For the first standard setup, reserve at least 13 GB on Apple Silicon Macs or 17 GB on CPU platforms such as Intel Mac and Windows. All four base packages need approximately 25 GB or 41 GB respectively. Actual space depends on existing downloads, temporary files and your work. Follow the app’s compatibility check.
An optional local transcription package for reference recordings is about 486 MB. It helps fill in the transcript, which still needs manual review. It is not required for every narration workflow.
A practical order for first setup
- Download the installer matching your system and chip, install it and open Shengyue Voice.
- Run the compatibility check on the activation screen, prepare the resources it requests and listen to fixed test samples.
- Check quality, speed and compatibility before purchasing. Email freeteamlite@gmail.com for purchase or setup questions.
- Send your device code and order details to your seller, then activate with the SY2 code issued for this device.
- Prepare the packages your work needs. Generate a short real script and review it before starting a long project.
What works offline, and what still needs the internet?
After the required runtime and packages are installed, narration, character voice locking, reference-voice reuse and license verification run locally. There is no online license check for each generation. Shengyue does not upload scripts, reference recordings or finished work. Adding packages, repairing missing components and downloading updates require an internet connection.
If an external AI agent writes scripts or calls the CLI, its own text handling and cloud use depend on its settings and workflow. Running narration locally does not make an external writing or reasoning service work offline.
Generation quotas and device-bound activation
Creation requires a purchased, activated license. There is no generation-count quota after activation, but text length, batch size, export duration and hardware limits still apply. Before activation, only compatibility checks and fixed test samples are available; you cannot freely edit scripts or export customer work.
SY2 activation is device-bound and can be verified offline. Price, device count, updates and transfer terms depend on your purchase agreement. Contact the original seller to reissue a code for a different computer. You can email freeteamlite@gmail.com with purchase or activation questions. This website does not currently offer online checkout.