vadepaysa

Hello HN,

I built https://inboxtutor.net , a tool (more like a toy) I use to send myself daily lessons over email, generated using Gemini.

ChatGPT and Perplexity offer something similar, but they require opening an app, and their lesson continuity is often weak. I kept getting repeated content. So I built something that works entirely through email. The whole experience lives in my inbox, a place I’m checking anyway.

How it works: 1. Enter what you want to learn (“Teach me Japanese for my trip”) 2. Verify your email 3. Receive personalized AI generated lessons in your inbox daily 4. Reply to any lesson to ask questions, take quizzes, or reshape the content

That’s it. No app. No dashboard. Just email.

You can also attach context using PDFs, URLs, or pasted text you want incorporated into your lessons.

I’ve found it a nice async way to learn and thought I’d share it here.

Try it: https://inboxtutor.net/

Hello HN!

I built OpenCodeSpace https://github.com/ngramai/opencodespace — a CLI tool that lets you launch disposable, self-hosted VS Code environments in one command.

It’s like a self-hosted Code Spaces, but:

- Works locally with Docker or remotely on Fly.io (AWS, GCE coming soon) - Auto-configured with Claude Code, OpenAI, Gemini CLI - Designed for YOLO mode development: temporary, parallel, throwaway sessions

As Claude Code and Gemini CLI became core to my dev workflow, I kept needing a way to run them in parallel, in clean, isolated environments in full YOLO mode (`--dangerously-skip-permissions`) without worrying about its effects on my computer.

I tired devcontainers, but not all repos have them configured and remote devcontainer setup is a pain. I also tried git work tree + multi window local VSCode, but I felt all the branching and folder replication on my local desktop a bit messy.

I wanted something simpler, so I built this.

How it works:

* Run `opencodespace .` in any folder * It checks for `.opencodespace` (or initializes one) * You choose: run locally with Docker or remotely on Fly.io * Browser-based VS Code opens up with everything ready

Install: pip install opencodespace

Code: https://github.com/ngramai/opencodespace

It’s still early days, but I love using this. I often spin up a few of these on Fly to let Claude code automatically chug away on bugs while I focus on heavier tasks in my local IDE.

Would love your feedback and ideas!

I find myself using QA on the codebase (not a generic LLM) more often. As a maintainer of internal and external facing python libs, I am wondering how much time I should spend maintaining working examples and docs.