One thing I've realized while building with AI is that context isn't a prompt problem anymore.
It's an engineering problem.
Over the last few weeks I've been using TokenCap on larger projects and found myself repeatedly asking:
- Why is the AI reading files that don't matter?
- Why am I sending the same context over and over?
- Why does every task start from scratch?
So v1.6.0 focuses on making context smaller, smarter, and more reusable.
What's new in v1.6.0
Smarter Context Retrieval
Only fetches the parts of your project that are actually relevant to the current task.
Improved Context Packing
Better prioritization means higher signal and fewer wasted tokens.
Better Dependency Resolution
Related files and components are grouped more intelligently before packing.
Faster Generation
Optimized pipeline for noticeably quicker context creation.
Bug fixes & stability improvements
Because reliable developer tooling matters more than flashy features.
Explore TokenCap
Website: https://tokencap.vansharora.app
npm: https://www.npmjs.com/package/tokencap
VS Code Marketplace: https://marketplace.visualstudio.com/items?itemName=VanshArora21.tokencap
Open VSX: https://open-vsx.org/extension/VanshArora21/tokencap
If you're building with AI, give it a try and let me know what you'd like to see in the next release.
Feedback, issues, and contributions are always welcome.
opensource #showdev #vscode #typescript #ai #contextengineering #developers #webdev #programming