Text to Speech FAQ
Practical answers about the current VoiceForge web service and its limits.
1. Is VoiceForge free to use?
Yes. VoiceForge is free to use without an account for normal web usage. Each TTS request supports up to 5,000 characters, and fair-use and rate limits apply so the public service can remain available.
2. How many voices and languages are available?
The current interface contains 27 configured neural voice profiles across 12 language families, including English, Chinese, Japanese, Korean, German, French, Spanish, Portuguese, Italian, Russian, Hindi, and Arabic.
3. What audio format does VoiceForge download?
The current web API and download flow provide MP3 output. WAV and OGG are not offered by the current deployment.
4. Does VoiceForge execute arbitrary SSML?
No. The current engine does not execute arbitrary custom SSML submitted through the web interface. The SSML compatibility helper validates and converts common markup into readable plain text before TTS generation. Global speed and pitch controls are still available.
5. How does automatic translation work?
When the selected voice language differs from the detected script language, VoiceForge can request a translation from a third-party translation service. You can edit the returned text before generating audio. Third-party provider policies apply to the translated request.
6. Is my text stored?
VoiceForge does not keep submitted text or generated audio as server-side history after processing. Generation history shown in the browser is stored locally on your device. Third-party providers may handle requests according to their own privacy policies.
7. Can I use generated audio commercially?
Commercial rights depend on the upstream speech service terms, your rights to the source text, and the specific project. Review the applicable third-party terms before commercial publication or large-scale use.
8. Why can generation be slow?
Generation time depends on text length, upstream response time, current server capacity, and whether the hosting environment needs to wake from inactivity. The service may return a rate-limit or capacity response when it is busy.