I’m an indie developer, and I wanted a reasonably priced, powerful setup for AI‑assisted coding that still keeps me in control of the work. After some experimentation, I’ve ended up with a roughly 60 USD/month stack that lives mostly inside VS Code and covers planning, coding, refactoring, and research.
The core stack
- GitHub Copilot Pro – Inline completions, small refactors, and “fill in the obvious code” while I’m in the editor. It’s my default background assistant that keeps momentum going.
- Claude Opus 4.6 – For higher‑level thinking: architecture discussions, planning the next steps, and non‑trivial refactors where I need long‑context reasoning and a more careful explanation.
- Gemini 3.1 Pro – For alternative solutions and explanations. I use it when I want a second opinion on an approach, or a clearer explanation of complex code or algorithms.
- Perplexity Pro (deep research) – For research and verification: comparing libraries, checking claims against multiple sources, and quickly collecting links I can inspect myself.
Open questions for you
- Do you also stack multiple AI tools for development, or do you prefer going all‑in on a single provider?
- How do you manage prompts and context across tools (especially inside an editor like VS Code) without drowning in windows and tabs?
- If you had to cut this down to two AI services, which ones would you keep and why?
Drafted with the help of an AI assistant, edited by me.
Top comments (0)