My Actual AI Stack for Building Apps in 2026 — All Free
Source: Dev.to
Introduction
If this is useful, a ❤️ helps others find it.
I’ve shipped 7 Mac apps in the past year, and every AI feature in them runs on free tools. Below is the exact stack I use, why I chose each component, and where the limits lie.
Gemini 2.5 Flash Preview via REST API
Cost: Free tier — 500 requests/day, no credit card required
Use for: Log diagnosis, document analysis, text classification, or any task that needs strong reasoning.
The free tier is genuinely sufficient for developer tools with intermittent AI use; I’ve never hit the daily limit in normal usage.
Get a key: (takes ~2 minutes).
Running Open‑Source LLMs Locally
Cost: Free, open source
Use for: Privacy‑sensitive processing, offline use, high‑volume tasks.
Install
brew install ollama
Pull a model
ollama pull gemma2
Run
ollama run gemma2
Models I actually use
| Model | Strength | Typical RAM |
|---|---|---|
gemma2 | Good general reasoning | 8 GB |
qwen2.5-coder:1.5b | Fast code autocomplete, tiny footprint |