llmcopy

pipx install llmcopy

uv tool install llmcopy

Either one gives you a llmcopy command that works from any folder and lives in its own environment, so it never mixes with your projects' packages. pipx is the long-standing tool for exactly this; uv is newer, faster, and fetches Python itself if the machine has none. pip install llmcopy works too, but installs into whatever Python pip belongs to. Python 3.9+, nothing else; Windows, macOS and Linux. Then run llmcopy in a project folder.

Copy a code project to the clipboard, ready to paste into ChatGPT, Claude, Gemini or any other AI chat. Glance at the tree, press Enter, paste.

• Knows what belongs to the project: asks git, and leaves out dependencies, build output, lock files, binaries. What is left out stays visible, dimmed, and can be ticked; i hides it.
• Counts tokens per file and per folder against a budget. When the project is too big for one message, the heaviest folders come first and the cursor is already on the biggest one: Space unticks it.
• Keyboard and mouse. Every key is listed in the bottom line, and the bottom line is clickable.
• s saves the selection as the default for this folder. a selects or clears everything. / filters by name.
• Files that look like they contain credentials are held back and marked.
• Starts in about 0.1 s. One file, no dependencies.

For LLMs

LLMS.md describes everything - behaviour, options, file formats, the structure of the code - in one document written to be read by an AI assistant. Give it to yours and ask.

MIT licence.