Improve your AI's performance by moving from a single, massive context file to a modular library of small, focused markdown files. This system allows the AI to intelligently load only the most relevant information for each task, enabling you to use simple, 'lazy' prompts and get better results.
with Claire Vo
Instead of one large claude.md file, create a dedicated folder (e.g., LLM Context) in a tool like Obsidian. Break down your knowledge into dozens of small, focused markdown files. For example, create separate files for brand_guidelines.md, marketing_channels.md, and personal_profile.md.
For broad categories like 'business,' create a main index file (e.g., business_profile.md). This file doesn't contain all the information itself; instead, it acts as a map. It should tell the AI where to find more specific details, like 'You can find my company overview here...' or 'Details about my course are in this file...'
In your global claude.md file, provide a simple instruction that directs the AI to the correct starting point. For example: 'If I ask you for help with something related to my business, use my business_profile.md file first.' This teaches Claude how to navigate your library and pull only the necessary context.
You don't need to build the entire library at once. At the end of a session with your AI, make a habit of asking it to help you document what it learned. This turns every interaction into an opportunity to refine your context files, making the system smarter over time.
Claude, what'd you learn today that we should document?
Build a custom research assistant that automatically finds relevant academic papers and generates high-quality summaries. This system delivers a daily research digest to your to-do list, helping you stay on top of your field with minimal effort.
Escape the friction of traditional GUI-based tools by building a bespoke task management system that runs in your terminal. This workflow uses Claude Code and markdown files to create and manage your to-do list with simple slash commands and natural language.
Leverage OpenAI Codex for high-level software architecture by using the `Plans.md` technique. This workflow guides the AI to generate a comprehensive, multi-step implementation plan for large projects before writing any code.
We've made ChatPRD affordable so everyone from engineers to founders to Chief Product Officers can benefit from an AI PM.