Read Claude Code's replies in your own language while the model still reasons in English for best quality.
Keep code, commands, and file paths untouched in translated replies so nothing technical gets mistranslated.
Toggle translation on or off with a simple file switch without uninstalling anything.
| partyfly/cc-lingua | 0xallam/my-recipe | 0xhassaan/nn-from-scratch | |
|---|---|---|---|
| Stars | 0 | — | 0 |
| Language | Python | Python | Python |
| Last pushed | — | 2022-11-22 | — |
| Maintenance | — | Dormant | — |
| Setup difficulty | easy | moderate | moderate |
| Complexity | — | 2/5 | 4/5 |
| Audience | developer | general | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires your own paid DeepSeek API key, no key is bundled with the project.
cc-lingua is a small add-on for Claude Code, Anthropic's command line coding assistant, that translates what appears on your screen into your own language while leaving the model's actual reasoning in English. The idea is that English gives the model its best reasoning and avoids extra processing cost for languages like Chinese or Japanese, but you as the user still get to read replies in your own language. It works as a hook, a small piece of code that Claude Code runs at certain points, called MessageDisplay. When Claude Code is about to show you a reply, cc-lingua intercepts the on-screen narration and sends it to DeepSeek, a separate paid translation service, to convert it into your chosen language before it reaches your screen. Code blocks, file paths, shell commands, URLs, and other technical content are left untouched, only the plain English explanation text gets translated. The saved conversation history and everything the model itself sees stay in English. This is described as bring your own key: there is no server, account, or subscription tied to cc-lingua itself. You need to buy your own API key from DeepSeek separately, which the README says typically costs a few cents a day for normal use, and identical text is cached locally so repeated phrases are not translated twice. The tool is designed to fail open, meaning that if the key is missing, the network fails, or anything else goes wrong, it simply shows the original English text rather than blocking or breaking your session. Installation is done by pasting an instruction into Claude Code itself, which then clones the repository and sets up the hook automatically. It can be turned on or off with a simple file toggle, and fully removed by deleting its folder. The project is released under the MIT license.
A Claude Code hook that translates only the on-screen reply text into your language via DeepSeek, while the model keeps reasoning in English, requiring your own DeepSeek API key.
Mainly Python. The stack also includes Python, Claude Code, DeepSeek API.
MIT license, use freely for any purpose including commercial use, as long as you keep the copyright notice.
Setup difficulty is rated easy, with roughly 5min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Verify against the repo before relying on details.