Your AI Is Only as Good as the Context You Give It

2026-07-26 · Source: Machine Learning on Medium · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering · Depth: Intermediate, quick

Summary

PackMD is a specialized tool designed to address the challenge of providing structured context to Large Language Models (LLMs) for improved performance. It evolved from an earlier tool, Git2txt, which simply mashed GitHub repository content into a single text file, leading to sloppy AI responses due to a lack of structure. PackMD leverages Markdown, which LLMs prefer due to its balance of simplicity and clarity, allowing models to efficiently use their context window and interpret content naturally, mirroring their training data. PackMD automatically detects the context type, whether a GitHub repository URL or a standard webpage URL. For GitHub repos, it bundles files into a single Markdown document with clear headers and language-specific code blocks. For webpages, it acts as a smart scraper, stripping noisy HTML and converting the core content into pure Markdown, thereby optimizing token usage and enhancing code generation quality.

Key takeaway

For AI engineers focused on improving LLM code generation and reducing token waste, you should prioritize providing structured context. Raw text or unformatted web content can lead to inefficient token usage and suboptimal AI responses. Consider adopting tools like PackMD to automatically convert source materials, such as GitHub repositories or API documentation, into clean, LLM-ready Markdown. This approach ensures your models receive context in their preferred format, leading to more accurate and useful outputs.

Key insights

LLMs perform best with structured context, especially Markdown, for efficiency.

Principles

• AI prefers Markdown for its balance of simplicity and clarity.
• Structured data improves LLM response quality over raw text.
• Markdown aligns with LLM training data, aiding natural interpretation.

Method

PackMD automatically detects context type (GitHub repo, webpage) and converts it into clean, LLM-ready Markdown, separating files with headers and language-specific code blocks.

In practice

• Use Markdown to structure code and documentation for LLMs.
• Convert noisy HTML to Markdown for efficient web content processing.
• Separate files with headers and code blocks in bundled context.

Topics

• Large Language Models
• Markdown Formatting
• Context Engineering
• Code Generation
• PackMD
• GitHub Integration
• Web Scraping

Best for: NLP Engineer, AI Engineer, Machine Learning Engineer, Software Engineer

Related on AIssential

• We Don’t Have an AI Problem. We Have a Context Problem. (And MCP Exposes It) — Unstructured context, not model limitations, is the primary bottleneck in scaling and maintaining LLM-based systems.
• Structured Language Model Generation with Outlines — Outlines ensures LLMs produce strictly structured outputs by masking invalid tokens during generation.
• How to Build a Scalable LLM-Powered README Generator — LLM context window limits for code analysis can be overcome by intelligent data selection.
• You’re Probably Wasting AI Tokens Without Realising It — Avoid making AI models rediscover information to significantly reduce token waste and improve response quality.
• My cheatsheet for a clean context — Effective coding agent use requires mindful context management, precise prompting, and systematic debugging.
• TokenMizer: Graph-Structured Session Memory for Long-Horizon LLM Context Management — LLM session history is a structured knowledge artifact, not flat text, enabling efficient context management.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Machine Learning on Medium.