I built a VS Code extension that gives AI instant context about your codebase - 20+ organic installs in 2 hours.

The problem

Every time you open ChatGPT or Claude, it has zero memory of your project.

You spend the first 10 minutes of every session re-explaining your stack, your folder structure, your naming conventions, your architecture decisions, before you can ask the actual question you came to ask.

63% of developers in the Stack Overflow 2025 survey (49,000 respondents) said AI tools lack crucial context about their codebase. This isn't a skill issue. It's a tooling gap.

What I built
DeepCtx - a free VS Code extension that auto-builds codebase context and injects it into any AI tool.

Here's how it works:

Install the extension — no account, no API key, nothing to set up

Open any project → pick your context slice.

Hit Scan & Build Context
Copy the generated .deepctx.md → paste into ChatGPT, Claude, or Cursor

Ask your question — the AI already knows your stack, your patterns, your file structure.

Changed-files mode — only rescans files modified since your last git commit. much faster for incremental updates.

Git metadata - every context file shows commit hash, branch name, and timestamp so the AI knows exactly how fresh the context is.

Confidence warnings - if large files were skipped during the scan, the context file explicitly warns the AI so it doesn't make assumptions about incomplete areas.

What makes it different
Unlike Cursor's .cursorrules (which you write manually and never update), DeepCtx auto-generates and auto-updates your context. And unlike native IDE AI tools, the output is portable — it works with any AI tool, not just the one you're coding in.

Try it
Search DeepCtx in VS Code extensions or install directly:
👉 [https://marketplace.visualstudio.com/items?itemName=deepctx.deepctx]

Completely free!

Would love honest feedback!