pypi / supertonic
supertonic
What changed in supertonic 1, from its official release notes, upgrade guides and advisories. Sources checked 2026-10-28.
2 breaking1 security9 new APIs and options1 behavior and defaults1 tooling1 fixes and other
See which of these apply to your code, ranked, in your agent:
npx @unstale/cli initsupertonic 115 changes
Breaking changes 2
Pre-flight middleware rejects POST /v1/styles/import requests exceeding 1 MiB before buffering the body.
The residual numpy<2.0 pin was removed from requirements.txt, making it consistent with the package specification.
Security 1
Starts a loopback-only local HTTP server for TTS; binding beyond loopback is opt-in and emits a stderr warning.
New APIs and options 9
Adds a health endpoint that returns readiness and loaded model information.
Adds an endpoint to retrieve built-in and imported custom voices.
Adds an OpenAI Audio Speech-compatible alias for speech synthesis.
Adds an endpoint that accepts multipart or JSON uploads of Voice Builder style JSON and persists them per model.
Adds an endpoint for native synthesis with the full parameter set.
Adds batch synthesis for up to 64 items in one JSON request, returning base64-encoded audio.
The server supports wav, flac, and ogg (Vorbis) response formats; MP3, AAC, and Opus are not shipped.
Overrides per-model custom-voice storage with a single shared directory.
Adds a Python package for mounting the FastAPI app inside a larger ASGI service.
Behavior and default changes 1
The environment variable now overrides the cache directory on every code path, and late-set values are honored through call-time resolution.
Tooling 1
The
[serve]extra installs the dependencies needed for the local HTTP server.
Fixes and other changes 1
Python 3.13 is officially supported, expanding the supported Python range.