Releases: LinuxUserLost/ai_script_rip
v0.3
First day relese, inside folder is .sh files linked to .py's. .sh works better for hotkeys. ai_prompter and ai_chat_cleaner both run as terminal and hotkeys, script_control is terminal only.
more to come later, the html document should have a decent breakdown, any other questions put ?? in terminal and it has a built in help descriptions. all these were made to run if you put script folder in same folder as your vault,
example /documents/vault then /documents/scripts
though it is fully editable and should even be able to access vaults on other partitions if you change target directory location in terminal. ultra light weight program ability to add more pluggins to the script later.
folder location /scripts/
terminal_cleaner.sh
hotkey_cleaner.sh
terminal_prompter.sh
hotkey_prompter.sh
state.json
ai_prompter.py
ai_chat_cleaner.py
script_control.py
folder location /scripts/docs
conf_template.json
script_control_companion.md
ai_prompter_companion.md
RECOVERY.md
ai_logging_system_docs.html
folder location /scripts/logs
debug.log
folder location scripts/conf
ai_prompter.json
ai_chat_cleaner.json
AI Logging
System
A zero-dependency, hotkey-driven pipeline that captures AI conversations, structures them with markers, and archives them as searchable Obsidian vault entries — built entirely in pure Python 3.
System Overview
What the system does, why it exists, and how the three scripts form a pipeline.
This system solves a specific problem: AI conversations are ephemeral. You type a message, get a response, and unless you manually save it somewhere useful, that knowledge is gone. The AI Logging System automates capture, structure, and archival — turning raw chat into a searchable, linked Obsidian knowledge base with minimal friction.
The entire pipeline is triggered by two hotkeys. One builds your clipboard before a session, one archives it after. Everything in between is automatic.
End-to-End Flow
Rule List #obsidian #python #module
I coded a more advanced module (will upload it soon if requested) but I automated a little bit didn't like not knowing if the exact stucture inside was what I wanted for future updates so I made a rule document and a first draft UI template. Template is currently in html and pops up in a browser but will the basis for a java ui window that I will attempt to mirror in a terminal UI that is easy to navigate console incase you cant have ktinker or need to debug/refresh UI which hopefully isn't a problem.
Some rules are slightly too tight and will need to be ammended for future changes but is enough to design the program. I know wlclip as clipboard needs to have some other options for other distro and future OS compatability.
*The gui is currently in html just for a quick template design, the rules have it coded for python so it communicates with rest of script and future modules better.