feat: add meeting transcription app scaffolding

- Add Python backend structure with FastAPI for transcription/summarization
- Add React UI with audio recording, transcript, and summary views
- Configure Tauri to manage Python backend lifecycle
- Set up Windows cross-compilation with cargo-xwin
- Add Gitea CI workflow for lint, test, and multi-platform builds
- Configure ESLint, Prettier, and Vitest for code quality

Note: App scaffolding only - Python env and models not yet set up

src/backend/models/llm.py

@@ -0,0 +1,67 @@
+"""Local LLM for meeting summarization using Llama."""
+
+from pathlib import Path
+from typing import Optional
+
+from llama_cpp import Llama
+
+
+class LlamaSummarizer:
+    """Handles meeting summarization using local Llama model."""
+
+    def __init__(self, model_dir: Path, model_size: str = "1B"):
+        """Initialize Llama model."""
+        self.model_dir = model_dir
+        self.is_loaded = False
+
+        model_path = model_dir / f"llama-3.2-{model_size}-instruct-Q4_K_M.gguf"
+
+        try:
+            self.llm = Llama(
+                model_path=str(model_path),
+                n_ctx=8192,  # Context window
+                n_threads=4,  # CPU threads
+                n_gpu_layers=-1,  # Use GPU if available
+                verbose=False,
+            )
+            self.is_loaded = True
+        except Exception as e:
+            print(f"Failed to load Llama model: {e}")
+            self.is_loaded = False
+
+    async def summarize(self, transcript: str) -> Optional[str]:
+        """Generate a meeting summary from transcript."""
+        if not self.is_loaded:
+            return None
+
+        prompt = f"""<|begin_of_text|><|start_header_id|>system<|end_header_id|>
+
+You are a helpful assistant that creates concise meeting summaries. Focus on:
+- Key decisions made
+- Action items and who owns them
+- Important discussions and their outcomes
+- Next steps
+
+Keep the summary structured and easy to scan.<|eot_id|><|start_header_id|>user<|end_header_id|>
+
+Please summarize this meeting transcript:
+
+{transcript}<|eot_id|><|start_header_id|>assistant<|end_header_id|>
+
+Meeting Summary:
+"""
+
+        try:
+            response = self.llm(
+                prompt,
+                max_tokens=1024,
+                temperature=0.7,
+                top_p=0.9,
+                stop=["<|eot_id|>", "<|end_of_text|>"],
+            )
+
+            return response["choices"][0]["text"].strip()
+
+        except Exception as e:
+            print(f"Summarization error: {e}")
+            return None