Parakeet (local, multilingual)
Parakeet (local, multilingual) is one of the components that act as the hearing of a voice agent: it turns what the caller says into text so the agent can understand it. It runs on the NappAI server: no credential, the audio is not sent to any outside provider and there is no per-minute cost. It understands 25 European languages, such as Spanish, English, French, German, Italian or Portuguese, and recognises on its own the language it hears. It does not understand Catalan, Galician or Basque. It is the recommended STT to start with.
It plugs into the STT connector of the Voice Agent node. It holds the speech recognition settings the agent uses on every call.
When to use it
- You are starting a voice agent and want it to hear without setting up any outside account.
- You do not want your customers’ audio to leave the NappAI server.
- You get calls in several European languages and want the agent to answer in the caller’s language.
Configuration
| Field | What it does |
|---|---|
| Language | The call’s main language, for example es. Default: es. |
| Detect Language | Parakeet always transcribes in the language it hears. When on, the agent also switches language with the caller (its voice and its answers), among those in Supported Languages. When off, the agent stays in the Language you set. Off by default. |
Advanced options
| Field | What it does |
|---|---|
| Supported Languages | Languages the agent can switch to. Default: es and en. Only used when Detect Language is on. |
Output
The STT output holds the speech recognition settings. It connects to the STT connector of the Voice Agent node, or to Fallback STT when you turn on Fallback Engines and want it as a backup. The editor will not let you plug it into a TTS connector by mistake.
Example
A seaside hotel gets calls in Spanish, English and German.
- Add Parakeet (local, multilingual) and connect it to the STT connector of the Voice Agent.
- Leave Language as
esand turn on Detect Language. - In Supported Languages, keep
esandenand addde. - Use a TTS that speaks those languages, such as Supertonic (local), and test in the playground.
Good practices
- If you turn on Detect Language, check that the connected TTS can speak every language in Supported Languages.
- For Catalan, Galician or Basque, use a cloud STT such as Azure Speech STT.
- The differences between local and cloud engines are explained in Quality, noise and latency.
Related components
- Voice Agent: the node this component connects to.
- Supertonic (local): the recommended local voice to go with it.
- Vosk (local): lighter local recognition, Spanish and English only.
- Azure Speech STT: cloud recognition that also understands Catalan, Galician and Basque.