For everyone who constantly creates new projects for little experiments, a one-file Ruby script to quickly manage and navigate to keep them somewhat organized

Ever find yourself with 50 directories named test, test2, new-test, actually-working-test, scattered across your filesystem? Or worse, just coding in /tmp and losing everything?

try is here for your beautifully chaotic mind.

Instantly navigate through all your experiment directories with:

• Fuzzy search that just works
• Smart sorting - recently used stuff bubbles to the top
• Auto-dating - creates directories like 2025-08-17-redis-experiment
• Zero config - just one Ruby file, no dependencies

curl -sL https://raw.githubusercontent.com/tobi/try/refs/heads/main/try.rb > ~/.local/try.rb

# Add to your shell (bash/zsh)
echo 'eval "$(~/.local/try.rb init ~/src/tries)"' >> ~/.zshrc

You're learning Redis. You create /tmp/redis-test. Then ~/Desktop/redis-actually. Then ~/projects/testing-redis-again. Three weeks later you can't find that brilliant connection pooling solution you wrote at 2am.

All your experiments in one place, with instant fuzzy search:

$ try pool
→ 2025-08-14-redis-connection-pool    2h, 18.5
  2025-08-03-thread-pool              3d, 12.1
  2025-07-22-db-pooling               2w, 8.3
  + Create new: pool

Type, arrow down, enter. You're there.

Not just substring matching - it's smart:

• rds matches redis-server
• connpool matches connection-pool
• Recent stuff scores higher
• Shorter names win on equal matches

• Shows how long ago you touched each project
• Recently accessed directories float to the top
• Perfect for "what was I working on yesterday?"

• Clean, minimal interface
• Highlights matches as you type
• Shows scores so you know why things are ranked
• Dark mode by default (because obviously)

• Everything lives in ~/src/tries (configurable via TRY_PATH)
• Auto-prefixes with dates: 2025-08-17-your-idea
• Skip the date prompt if you already typed a name

Add to your ~/.bashrc or ~/.zshrc:

# default is ~/src/tries
eval "$(~/.local/try.rb init)"

Or if you want to customize the location:

eval "$(~/.local/try.rb init ~/src/tries)"

try                 # Browse all experiments
try redis           # Jump to redis experiment or create new
try new api         # Start with "2025-08-17-new-api"
try --help          # See all options

• ↑/↓ or Ctrl-P/N - Navigate
• Enter - Select or create
• Backspace - Delete character
• ESC - Cancel
• Just type to filter

Set TRY_PATH to change where experiments are stored:

export TRY_PATH=~/code/sketches

Default: ~/src/tries

• One file, no dependencies
• Works on any system with Ruby (macOS has it built-in)
• Fast enough for thousands of directories
• Easy to hack on

Your brain doesn't work in neat folders. You have ideas, you try things, you context-switch like a caffeinated squirrel. This tool embraces that.

Every experiment gets a home. Every home is instantly findable. Your 2am coding sessions are no longer lost to the void.

Q: Why not just use cd and ls? A: Because you have 200 directories and can't remember if you called it test-redis, redis-test, or new-redis-thing.

Q: Why not use fzf? A: fzf is great for files. This is specifically for project directories, with time-awareness and auto-creation built in.

Q: Can I use this for real projects? A: You can, but it's designed for experiments. Real projects deserve real names in real locations.

Q: What if I have thousands of experiments? A: First, welcome to the club. Second, it handles it fine - the scoring algorithm ensures relevant stuff stays on top.

It's one file. If you want to change something, just edit it. Send a PR if you think others would like it too.

MIT - Do whatever you want with it.

Built for developers with ADHD by developers with ADHD.

Your experiments deserve a home. 🏠