Switch from OpenAI to local Ollama to avoid quota issues

manifests/dot-ai-stack/values.yaml

@@ -4,7 +4,13 @@ dot-ai:
 
   ai:
     provider: openai
-    model: "gpt-5.1"  # Using GPT-5.1 from your available models
+    model: "llama3.1:70b"  # Using local Ollama model
+    customEndpoint:
+      enabled: true
+      baseURL: "http://ollama.ai-stack.svc.cluster.local:11434/v1"  # Point to Ollama service
+      embeddingsBaseURL: "http://ollama.ai-stack.svc.cluster.local:11434/v1"
+      embeddingsModel: "nomic-embed-text"  # Ollama embedding model
+      embeddingsDimensions: "768"  # Nomic embeddings dimension
 
   secrets:
     auth: