tvmux 0.5.3

📺 tvmux

Per-project/window tmux recorder using asciinema; records the current pane
and follows the user around the current window.

🎥 Usage

Install tvmux from pip or uv, or run standalone from uvx, like so:

$ uvx tvmux --help

Or, if installed, run from inside a tmux pane:

# Start recording
tvmux rec
# list ongoing recordings
tvmux rec ls
# stop them all, like you would a Docker container
tvmux rec stop $(tvmux ls -q)

By default, it'll save to ~/Videos/tmux/YYYY-MM/. See all configuration options:

# Show default config (TOML format)
tvmux config defaults

# Show available environment variables
tvmux config defaults --format=env

# Show your current config
tvmux config show

To customize, create ~/.tvmux.conf or use environment variables like TVMUX_OUTPUT_DIRECTORY.

🔗 links

• 🏠 home
• 📚 pydoc
• 📚 api
• 🐱 github
• 🐍 pypi

🌍 See also

📺 asciinema	The terminal recorder
🪟 textual	TUI library for Python
🗔 bittty	My terminal
🎬 sh2mp4	Convert this to MP4 files

TODO

Prep for next steps

• Need asciinema scrubber using bittty (wip)
• Start a basic TUI in Textual

tvmux 0.5.2

📺 tvmux

asciinema tmux recorder.

Python rewrite!

🎥 Why?!

So I can record tmux sessions and archive my workflow, then later use them for
AI training.

✋ How?

$ uvx tvmux --help

🔗 links

• 🏠 home
• 📚 pydoc
• 🐱 github
• 🐍 pypi

🌍 See also

📺 asciinema	The terminal recorder
🪟 textual	TUI library for Python
🗔 bittty	My terminal
🎬 sh2mp4	Convert this to MP4 files

TODO

Prep for next steps

• Need asciinema scrubber using bittty (wip)
• Start a basic TUI in Textual

tvmux 0.5.1

📺 tvmux

asciinema tmux recorder.

Python rewrite!

🎥 Why?!

So I can record tmux sessions and archive my workflow, then later use them for
AI training.

✋ How?

$ uvx tvmux --help

🔗 links

• 🏠 home
• 📚 pydoc
• 🐱 github
• 🐍 pypi

🌍 See also

📺 asciinema	The terminal recorder
🪟 textual	TUI library for Python
🗔 bittty	My terminal
🎬 sh2mp4	Convert this to MP4 files

TODO

Prep for next steps

• Need asciinema scrubber using bittty (wip)
• Start a basic TUI in Textual

tvmux 0.5.0

📺 tvmux

asciinema tmux recorder.

Python rewrite!

🎥 Why?!

So I can record tmux sessions and archive my workflow, then later use them for
AI training.

✋ How?

$ uvx tvmux --help

🔗 links

• 🏠 home
• 📚 pydoc
• 🐱 github
• 🐍 pypi

🌍 See also

📺 asciinema	The terminal recorder
🪟 textual	TUI library for Python
🗔 bittty	My terminal
🎬 sh2mp4	Convert this to MP4 files

TODO

Prep for next steps

• Need asciinema scrubber using bittty (wip)
• Start a basic TUI in Textual

tvmux 0.4.1

📺 tvmux

asciinema tmux recorder.

Python rewrite!

🎥 Why?!

So I can record tmux sessions and archive my workflow, then later use them for
AI training.

✋ How?

$ uvx tvmux --help

🔗 links

• 🏠 home
• 📚 pydoc
• 🐱 github
• 🐍 pypi

🌍 See also

📺 asciinema	The terminal recorder
🪟 textual	TUI library for Python
🗔 bittty	My terminal
🎬 sh2mp4	Convert this to MP4 files

tvmux 0.4.0

📺 tvmux

asciinema tmux recorder.

Python rewrite!

🎥 Why?!

So I can record tmux sessions and archive my workflow, then later use them for
AI training.

✋ How?

$ uvx tvmux --help

🔗 links

• 🏠 home
• 📚 pydoc
• 🐱 github
• 🐍 pypi

🌍 See also

📺 asciinema	The terminal recorder
🪟 textual	TUI library for Python
🗔 bittty	My terminal
🎬 sh2mp4	Convert this to MP4 files

tvmux 0.3.1

📺 tvmux

So I want to capture my dev work in tmux so I can train models on how to program
and actually use a computer. More data is good.

This is my current tmux setup, including my tmux.conf.

0.3.1

This shell script records a session following the active pane. After some turd
polish and more testing, I plan to expand it to work with one window per
project, quantizing recordings, asciinema streaming etc etc

To-do

General

• get the monitor panel working so I don't need to manually create it each
  reboot.
• Write a ./configure script that checks all the things we need.
• make build and make install to $PREFIX, defaulting to ~/.local/

0.4

• Record everything and selectively switch between buffers
  • timestamp logs with an awk script
• Get alternative buffers working
• Break stuff into a library + pack it
• quantize because spinners are a menace
  • thought: set a cps limit and send full buffers if exceeded

__future__

• Command line args / flags / config
  • Consider a generic settings approach that works with config file, args
    and env vars as first class citizens.
• Move deployment/install code into the script itself
  • Put markers in so it can unpack itself to source 😎
  • Think about how to pack docs, unpacking full source with makefile and
    README.md? 🤯