Blog
The notes here focus on practical implementation choices, operational tradeoffs, and small-team constraints.
LeeCut is a macOS desktop workflow for product short video editing, connecting material import, product knowledge, AI copy, voiceover, timeline editing, MP4 export, and CapCut/Jianying draft export.
The Root and Affix Trainer is a LeewoAI English-learning tool built around 306 frequency-audited prefixes, roots, and suffixes, combining question modes, spaced review, confusion groups, mnemonic images, pronunciation, and local progress.
A practical overview of LeeMate based on the current project structure, covering released features, suitable use cases, workflow tips, and functional boundaries.
A detailed guide based on LeeSnap's current code structure, covering setup, shortcuts, screenshot annotation, selection translation, AI configuration, clipboard workflows, reminders, and mouse enhancement.
A working view of how small teams can design AI-assisted support systems that reduce response time without losing control of escalation, context, or quality.
Useful automation is not defined by how much language a model can generate, but by whether a workflow becomes clearer, faster, and easier to operate after the model is inserted.
Small AI products usually do better with simple infrastructure, explicit observability, and a bias toward maintainable operations over impressive architecture.