Stack
Every frame, and exactly where you are.
Visual debugger · Python 3.14+
See your program as it really runs.
yathaavat is a visual debugger for Python that lives in your terminal — stack, source, locals, exceptions, and asyncio tasks in one cockpit. Launch it, connect to it, or safely attach. Keyboard-first, at home over SSH.
$curl -fsSL https://yathaavat.pages.dev/install | sh
Source
Live gutter markers — ● ▶ — and your current line.
Locals & watches
Real values, expanded inline, tracked as they change.
Console
Evaluate any expression, with tab-completion.
The hard parts, made legible
The states that usually send you back to print() — exceptions and async — rendered so you can actually read them.
Exception stop
Exceptions, read like a story
On every exception stop, yathaavat draws the traceback as a tree — chained causes labelled ↳ caused by, full ExceptionGroup support, the failing frame highlighted.
Enter jump to frame y copy a break at raise
Every value, no guessing
Expand locals and globals inline, watch expressions across steps, and evaluate anything in a console with real DAP tab-completion. Edit a value in place and keep going.
Enter expand e edit Ctrl+W watch g scope
Locals · expanded inline
How you steer
Drive execution at the speed of thought
Stepping is a transport control, not a menu. Your hands never leave the keyboard — continue, step, run-to-cursor, and you're three frames deeper before a mouse would have found the button.
▶cContinue
↷nStep over
↳sStep in
↰uStep out
⏎EnterRun to cursor
⌖Ctrl+EJump to line
Sessions
Ctrl+RLaunch a script
Ctrl+KConnect to a server
Ctrl+AAttach to a process
Ctrl+PCommand palette
Inspect
bToggle breakpoint
Ctrl+FFind in source
Ctrl+GGo to line
F2Zoom a pane
Three ways into a running program
However the code is running, yathaavat can reach it — with fuzzy discovery and history at every door.
Ctrl+R launch
Launch a script
Start any Python file under debugpy. Fuzzy file discovery and ~ expansion get you running in two keystrokes.
Ctrl+K connect
Connect to a server
Attach to a debugpy endpoint by host and port. The picker scans for live servers so you rarely type an address.
Ctrl+A attach
Attach to a process
Attach to a live PID. On Python 3.14, attach safely to a process that never imported debugpy.
Why
The terminal you already live in
The truth of a running program shouldn't cost you a GUI you can't SSH into, or a litter of print() you'll forget to delete. yathaavat renders the whole runtime where you already work — no browser, no daemon, no ceremony.
- Runtime
- Python 3.14+ on CPython, via debugpy and the Debug Adapter Protocol
- Ways in
- Launch, connect, attach, and safe attach on 3.14
- Interface
- A Textual TUI — keyboard-first, mouse-optional
- At home in
- SSH sessions, tmux panes, the terminal you trust
- License
- MIT — install with
uv, no account required
Questions, answered
Do I have to change my code? +
No. Launch any script as-is, connect to a debugpy server, or on Python 3.14 attach to a process that never imported debugpy via safe attach.
Does it work over SSH and tmux? +
Yes. yathaavat is keyboard-first and mouse-optional, designed from the start for remote sessions and split panes.
What do I need to run it? +
Python 3.14+ and uv. Install in one line — no PyPI account, no browser, no build step.
How is this different from pdb or an IDE debugger? +
pdb shows one line at a time; an IDE debugger needs the IDE. yathaavat gives you the whole frame at once, in the terminal.
Is it actually open source? +
MIT-licensed, end to end. Install with uvx --from git+https://github.com/indrasvat/yathaavat yathaavat.
Nothing hidden. Nothing assumed.
Install yathaavat and watch your next breakpoint exactly as it is.
$curl -fsSL https://yathaavat.pages.dev/install | sh