Get started with NorSpeak
NorSpeak is a browser workbench for text-to-speech, speech-to-text, and realtime voice calls. Use it on the page for free. Add an API key only when your own product needs to call the service.
Text to speech
Open the workbench, choose TTS, paste a script, and pick a supported language and voice. Play the result in the page, then download the same audio file the service returned — exporting does not synthesize a second time.
Use this for narration drafts, pronunciation checks, and short voiceovers. The format (MP3, WAV, OGG, or WebM) matches whatever the API actually produced.
Speech to text
Switch to STT to record, dictate, or upload audio. Review the transcript while the source is still at hand, correct names, and export clean text.
This mode is built for interviews, notes, and captions — a result you can take into another document, not a demo locked in the page.
Realtime conversation
Live starts a two-way voice session in the browser. Allow the microphone, connect, and speak. Captions accumulate in a timeline, and the orb shows listening, thinking, speaking, and interruption states.
Full duplex is the default for a natural exchange; push-to-talk helps in noisy rooms. When you finish, export the transcript and any recording that actually exists.
Free on the page, paid in your product
Trying TTS, STT, and Live in the NorSpeak workbench does not require a paid plan. Keys and credit are for servers, workers, automations, or apps that call the API outside this site.
Keep secrets on a server or in a Worker secret. Review usage from your account when you start making external requests.