How do you give your AI coding assistant context about your project architecture?
Source: Dev.to
Overview
Every time I open a new Claude Code or Cursor session, I spend the first 10 minutes explaining my project.
- “We use PostgreSQL because…”
- “Auth is JWT with 24 h expiry…”
- “Please follow REST conventions…”
How do you handle this?
- Write a
CLAUDE.mdmanually - Paste context every session
- Use a specific prompt template
- Something else entirely
I got tired of repeating myself and built LORE — an MCP server that reads your codebase and extracts these decisions automatically.
But I’m curious what the community does. What’s your approach?