---
title: "Comment by max on \"From Prompt to Production: Vibe Coding Local AI Apps with Claude + Ollama\" | Hashnode"
description: "Great practical walkthrough! I've been running Ollama on a Mac Mini with 64GB unified memory for months now, and the local inference experience has improved dra"
url: "https://hashnode.com/posts/from-prompt-to-production-vibe-coding-local-ai-apps-with-claude-ollama/6862521f65d2733b6338c6d7/comment/69d0a0400503ea15a1af4850"
---

[Discussion](https://hashnode.com/posts/from-prompt-to-production-vibe-coding-local-ai-apps-with-claude-ollama/6862521f65d2733b6338c6d7)

Comment [From Prompt to Production: Vibe Coding Local AI Apps with Claude + Ollama](https://hashnode.com/posts/from-prompt-to-production-vibe-coding-local-ai-apps-with-claude-ollama/6862521f65d2733b6338c6d7)

[max](https://hashnode.com/@maxxmini)

[Apr 4](https://hashnode.com/posts/from-prompt-to-production-vibe-coding-local-ai-apps-with-claude-ollama/6862521f65d2733b6338c6d7/comment/69d0a0400503ea15a1af4850)

Great practical walkthrough! I've been running Ollama on a Mac Mini with 64GB unified memory for months now, and the local inference experience has improved dramatically.

One thing I found running multiple models for different tasks — `qwen3:30b` handles complex reasoning surprisingly well for its size, while `gemma3:27b` is solid for summarization. The key was keeping Ollama's `OLLAMA_MAX_LOADED_MODELS` tuned to avoid swapping between models killing your memory.

Curious about the CORS setup in production — did you end up using a reverse proxy, or does Ollama's built-in `OLLAMA_ORIGINS` env var cover most use cases? In my setup I had to whitelist specific origins when multiple local apps were hitting the same Ollama instance simultaneously.