diff --git a/articles/en/community/configuring-spanish-deepgram-voice-agent-1408448594763059529.md b/articles/en/community/configuring-spanish-deepgram-voice-agent-1408448594763059529.md new file mode 100644 index 00000000..ec9dbfcc --- /dev/null +++ b/articles/en/community/configuring-spanish-deepgram-voice-agent-1408448594763059529.md @@ -0,0 +1,53 @@ +# Configuring Spanish Language in Deepgram Voice Agent + +When utilizing Deepgram's Voice Agent API, configuring language settings is crucial to ensure seamless operation, especially when dealing with non-English languages like Spanish. Here's a guide to properly configure your Voice Agent to support Spanish. + +### Overview + +Deepgram Voice Agent API allows communication functionalities using various languages. While the Nova-3 model is widely used, currently, it supports monolingual models for English, Danish, Dutch, German, and Swedish. For Spanish, it's essential to configure the system to use Nova-2 or another compatible model. + +### Configuration Setup + +To set up your Deepgram Voice Agent for Spanish language processing, here is how you should configure your agent: + +```json +{ + "agent": { + "language": "es", + "speak": { + "provider": { + "type": "deepgram", + "model": "aura-2-estrella-es" + } + }, + "listen": { + "provider": { + "type": "deepgram", + "model": "nova-2" + } + }, + "think": { + "provider": { + "type": "anthropic", + "model": "claude-3-5-haiku-latest" + }, + "prompt": "YOUR_PROMPT" + } + } +} +``` + +### Key Considerations + +- **Listen Provider**: When setting the listen provider for non-English languages like Spanish, ensure to use "nova-2" model, which currently offers better support for processing Spanish language. +- **Speak Provider**: The "aura-2-estrella-es" model is suitable for Spanish in the Voice Agent setup. +- **Think Provider**: You can continue using Anthropics' model fit explicitly for your cognitive processing needs (e.g., Claude models). + +### Conclusion + +Configuring Deepgram's Voice Agent API for Spanish involves selecting models that cater to the language-specific requirements. For Spanish, using the Nova-2 model along with the "aura-2-estrella-es" for speaking ensures proper handling of language nuances. If issues persist or the system behavior seems inconsistent, reach out to your Deepgram support representative (if you have one) or visit our community for assistance: [https://discord.gg/deepgram](https://discord.gg/deepgram) + +### References + +- [Deepgram Models and Languages Overview](https://developers.deepgram.com/docs/models-languages-overview#nova-3) +- [Deepgram Discord Community](https://discord.gg/deepgram) \ No newline at end of file