How I bundle my codebase so ChatGPT can actually understand it

Source: Dev.to

Problem

LLM chat apps are great at answering questions—until you point them at a real codebase.
When a project grows past a certain size, context becomes the bottleneck: too many files, too much noise, not enough structure.

Typical LLM chat apps work best when:

• Context is linear
• Files are grouped by meaning, not size
• References are precise (file + line)

Real repositories, however, are:

• Hierarchical
• Noisy
• Full of things the model doesn’t need

The result is vague or hallucinated answers.

Solution: srcpack

Instead of feeding the whole repo, generate an LLM‑optimized snapshot:

• Code is bundled by domain (e.g., web, api, docs, etc.)
• Each bundle is indexed with file paths + line numbers
• .gitignore is respected by default
• No configuration needed to get started

The output consists of plain‑text files, easy to upload or attach to ChatGPT, Gemini, Claude, etc.
The tool is called srcpack.

Workflows where srcpack shines

• Exploring large repos – Ask “where does auth actually live?” or “what touches billing?”
• Avoiding context limits – Instead of pasting files manually, attach a focused bundle.
• Sharing with non‑technical teammates – Upload an LLM‑friendly snapshot to Google Drive and share it.

Typical questions from PMs or designers:

• “What shipped this week?”
• “What’s still in progress?”
• “Which parts are risky?”

srcpack acts like a lightweight, read‑only AI interface to the codebase.

Usage

npx srcpack   # or bunx srcpack

Zero configuration is required by default.

Documentation & Source

• Docs:
• GitHub:

Closing Thought

I don’t think this is the final answer to LLM + codebase interaction, but it’s been a very practical improvement for day‑to‑day work. I’m curious how others are handling large‑repo context with LLMs—especially in fast‑moving projects.