Stop pasting 5,000 lines of logs into Claude. Use a secure context tunnel instead.
Tags: #ai #productivity #cli #security
The Problem: The "Wall of Text" Friction
We've all been there. You're debugging a nasty crash. You have a 2MB log file. You try to paste it into ChatGPT or Claude.
- ❌ The UI freezes.
- ❌ The text gets truncated.
- ❌ You realize you just pasted your API keys into a cloud chat history.
The Solution: vnsh (Vanish)
I built an open-source tool called vnsh. Think of it as an ephemeral "Dropbox" designed specifically for AI agents.
How it works:
- You pipe data in your terminal:
cat error.log | vn - It encrypts it locally (AES-256-CBC).
- It gives you a secure link.
- You give that link to Claude.
Because I built a native Model Context Protocol (MCP) server for it, Claude can actually "see" inside the encrypted link and read the file directly.
Quick Start
If you are on Mac/Linux (Homebrew):
brew tap raullenchai/vnsh
brew install vnsh
Or via NPM (Node.js):
npm install -g vnsh-cli
The "Magic" Workflow
Next time you have a git diff that is too long to explain:
git diff | vn
# Output: [https://vnsh.dev/v/abc...#k=](https://vnsh.dev/v/abc...#k=)...
Paste that URL to Claude. It stays fast, the server (me) can't read your code, and the data self-destructs in 24 hours.
Self-Hosting
Since it deals with sensitive data, I made it host-blind (the decryption key is in the URL hash fragment, never sent to the server). But if you are paranoid (like me), you can self-host the whole stack on your own Cloudflare account.
Check it out on GitHub: https://github.com/raullenchai/vnsh
Top comments (0)