German speaking LLM for language practice
Check local LLM options for German language practice and define a system prompt that provides corrections.
Check local LLM options for German language practice.
Summary
An enthusiast is exploring how to build a German‑speaking LLM that can correct spoken German. The post clarifies that speech‑to‑text is not a concern, and the focus is on having the model respond in German and provide corrections defined in a system prompt. The author explicitly wants more than a translation tool; they need interactive feedback that corrects pronunciation and grammar. They ask whether anyone has set up a local LLM for this purpose and what success has been achieved.
The discussion highlights the need for a teacher‑like interaction rather than a simple translation engine. No specific model or framework is mentioned, leaving the implementation open. The post invites community experience with local LLM setups for language practice.
Key changes
- User wants German speaking LLM for practice
- Speech‑to‑text not required
- Focus on corrections via system prompt
- LLM should not just translate
- No existing solution discussed