Guide
Local AI Starter Stack
Build a local AI starter stack with model runtimes, chat interfaces, coding assistants, vector search, and evaluation habits.
Steps
- Use Ollama or LM Studio as the model runtime.
- Add Open WebUI for a familiar local chat interface.
- Use Continue for coding assistance inside your editor.
- Add Qdrant, Chroma, or built-in retrieval when documents become important.
- Track model, prompt, latency, and quality notes so you can compare changes.
FAQ
What is the minimum local AI stack?
A runtime plus one model is enough to start. Add chat, coding, retrieval, and evaluation tools only when your workflow needs them.
Should beginners start with the largest model?
No. Start with a small model that runs reliably, then move up when you understand the speed and memory trade-offs.
Next steps
Use the model and tool directories to choose the concrete pieces for your local AI stack. Sponsor and affiliate placements will be added later.