~/ldgweb.com

ldgweb

Full stack developer. I'm building my own way of working with AI agents, one tool at a time: they read the context, they propose, I decide.

how I work

  1. Everything I learn goes into one memory: the docs of every project, my command toolbox, ideas and drafts. Agents read it before they start. beynt
  2. I code and run my servers with agents in the terminal, several sessions at once. edizium
  3. Every change an agent makes gets my review before it stays. agent-review-studio
  4. Quick notes, handwriting and throwaway text stay one shortcut away. appunto

I use these tools every day, and they change as my way of working does.

working on

Work with any AI. Keep what works.

The memory of all my work: the docs of every project, my toolbox of commands, ideas and drafts. Every AI app I connect over MCP reads it before it starts and proposes changes; I approve what stays. Open to anyone who builds with AI.

edizium

Where I code and run my servers with AI agents. Agent sessions side by side, SSH servers, guided runs and deploys across several machines. An agent can read a session and propose a command; I'm the one who runs it.

  • Rust
  • Tauri
  • in daily use, not public yet

open source

Tools from my own workflow, free to use.

agent-review-studio

How I review code written by agents. The agent's session sits next to the diff, and I decide file by file what to keep. Desktop app for Windows and the ars CLI.

  • Rust
  • TypeScript
  • MIT

appunto

For quick notes: text, handwriting, and a scratch page that is never saved. One shortcut shows it from any app, and everything stays on my machine. Windows, macOS and Linux.

  • Rust
  • TypeScript
  • MIT

all repositories →

contact