An offline alternative to Google Cloud Text-to-Speech
Google Cloud Text-to-Speech, Amazon Polly and Azure AI Speech are cloud APIs: every request sends your text to the provider, usage beyond a free tier is billed by the character, and you need an account, a key and an internet connection. SuperClonica offers the same kind of building block that runs on your own machine.
This page compares the two setups honestly, says when a cloud API is the better choice, and shows how to swap a cloud call for a local one.
Cloud API vs a voice you run
Cloud APIs are easier to start and scale to any load; a local engine trades that for privacy and a fixed cost.
- Cloud TTS: managed servers, many languages and voices, billing by usage, your text sent to the provider.
- SuperClonica: a Docker image with an HTTP endpoint, a Python library and a command-line tool, all running locally.
- After the one-time model download, about 385 MB, the server needs no internet.
- Voices are JSON files: the ten built-in voices, catalog voices, or a voice trained from your own recording.
When a cloud API is the better choice
Choose a cloud API when you need Chinese or a very long list of languages and regional variants, SSML control, managed scaling with service guarantees, or simply no server to maintain. SuperClonica speaks 31 languages, has no SSML and runs on hardware you manage.
When SuperClonica fits better
In our tests a short phrase took 0.12 s on an RTX 5070 Ti and 0.43 s on four threads of a Ryzen 7 5800X.
- Your texts are private: medical notes, internal documents, unreleased game dialogue.
- You need to work offline or inside a closed network.
- Your volume is large or unpredictable, and a per-character bill would grow with it.
- You want character voices, a brand voice trained from your own speaker, and the same voice in all 31 languages.
Swap the API call
Start the server on any CPU: docker run -d --name tts -p 8000:8000 -v superclonica-models:/models superclonica/superclonica:cpu. When GET http://localhost:8000/health answers, send a request: curl -X POST http://localhost:8000/synthesize -H "Content-Type: application/json" -d '{"text": "Hello from my own server!", "voice": "F1", "lang": "en"}' -o hello.wav. POST /synthesize takes text, voice, lang, seed and normalize and answers with a 16-bit mono WAV file. Set SUPERCLONICA_API_KEY to require an Authorization header, and keep the server behind HTTPS or on a private network. A custom voice goes into the request as its JSON object.
Costs and licenses
The SuperClonica engine and the Docker image are released under the Apache License 2.0. The Supertonic 3 model is under OpenRAIL-M, a royalty-free license that allows commercial use within its list of prohibited uses. The built-in voices and several catalog voices are free; other catalog voices are bought once. Check the cloud providers’ current prices on their own sites before you compare.