A developer posted a Show HN on August 7th for kivun-terminal-wsl, an open-source project aimed at getting Claude Code to properly render Hebrew, Arabic, and Persian text in the terminal. The submission landed with six points on Hacker News and zero comments — a quiet launch, but it pokes at a real gap in how AI coding tools handle non-Latin scripts. The repo name itself tells part of the story. "Kivun" is Hebrew for direction, an explicit nod to the right-to-left text flow these three languages share. The "-wsl" suffix points at Windows Subsystem for Linux as a core target — and that's precisely where terminal RTL support tends to break down in practice. This matters more than it might seem at first glance. Terminal emulators have historically been terrible at bidi text, often scrambling Hebrew or Arabic output into jumbled glyph order when mixed with code, prompts, or log lines. For anyone using Claude Code as a daily driver while working in those languages, garbled AI responses aren't just cosmetic — they make the tool effectively unusable for real work. What we can't tell you yet is how this project actually pulls it off. The source material retrieved for this story came through corrupted and unreadable, so specifics on implementation — whether it's a font configuration layer, a custom pager, WSL-specific patches, or something else entirely — aren't available from the post itself. Anyone evaluating it will want to check the GitHub README directly before assuming what's under the hood.
Key Takeaways
- kivun-terminal-wsl is an open-source project posted as Show HN on August 7th, targeting Hebrew, Arabic, and Persian rendering in Claude Code's terminal output.
- The repo name suggests a focus on right-to-left directionality ("kivun") with Windows Subsystem for Linux support implied by the "-wsl" suffix.
- Implementation details remain unconfirmed — the source content was corrupted during retrieval, so no claims about how the tool works should be taken at face value yet.
The Bottom Line
RTL rendering in terminals is a long-neglected corner of developer tooling, and it's good to see someone shipping for it instead of waiting on emulator maintainers. But with zero comments and unverifiable implementation details, this one needs a closer look at the repo before anyone bets their workflow on it.