About TreeSnap, well it simplifies this preparation of sharing code with Chat LLMs. In my use case, making o3 via ChatGPT a lot more cost-effective to use for these tasks compared to Cursor:
How it works:
Select files/folders → generates a single text bundle + ASCII tree
Instantly shows per-file and total token counts (list & treemap views)
Quickly trim unnecessary files or folders to match context limits
Fully offline and private (Tauri desktop app)
Also available as a free browser app—offline and private
Built in ~3 weeks, I use it daily to aid my workflow and reduce my expenses by minimizing the need to rely on Cursor's pricier model usage. I still rely on Cursor with grok 3 or gemini pro 2.5 pro mostly without getting crazy with usage fees. Feedback welcome—especially UX, performance, and pricing.