Claude-ts – Translation proxy to fix non-English token waste in Claude

Name: Claude-ts – Translation proxy to fix non-English token waste in Claude
Availability: InStock
Author: kiimdonglin

by kiimdonglin·Feb 22, 2026·4 points·0 comments

Visit Project View on HN

AI Analysis

●●SolidSolve My ProblemBig Brain

Fixes multilingual token waste by translating to English before Claude, not after.

Strengths

•Targets a real, quantifiable problem: non-English text wastes 2-3x tokens and degrades reasoning.
•Clever architecture: translation layer via cheap Haiku/Ollama costs almost nothing vs. the main Claude call.
•Local Ollama option eliminates translation costs entirely for privacy-conscious users.

Weaknesses

•Depends entirely on Claude Code CLI adoption; won't work if Anthropic changes the interface.
•Translation quality between 8 languages via Haiku is unproven; hallucinations could corrupt intent.

Post Description

When you use Claude Code in Korean, Japanese, or any non-English language, two things happen:

1. You waste tokens — non-English text takes 2-3x more tokens than English for the same meaning. Every prompt, every response, every turn in context is inflated.

2. Claude reasons worse — it spends context budget on language switching instead of actually thinking about your code.

I built claude-ts to fix this. It's a translation proxy that sits in front of Claude Code:

You (any language) → Haiku/Ollama (→ EN) → Claude Code (EN) → Haiku/Ollama (→ your lang) → You

Claude Code always works in English internally — better reasoning, fewer tokens. The translation costs almost nothing (Haiku) or literally nothing (local Ollama).

pip install claude-ts

- 8 languages supported (ko, ja, zh, th, hi, ar, bn, ru) - Free local translation via Ollama - Real-time agent tree visualization - All Claude Code features preserved