How do I install the Translate skill?

Run npx skillmds@latest add ryanmaclean/translate in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Translate safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Translate?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Translate free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Translate?

ryanmaclean (@ryanmaclean) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Translate by ryanmaclean — Agent Skill

Name: Translate
Author: ryanmaclean

Translate

Local multilingual translation via Ollama — 119 languages, no API key, no network, no data leaves the machine. Use when translating to/from Japanese, Chinese, Korean, French, Spanish, German, Portuguese, Russian, Arabic, or any other language. Also for localization, i18n text, and roundtrip validation. Qwen3:32b on Apple Silicon — faster and cheaper than cloud APIs.

ryanmaclean 3bc8e52 3 files · 8.6 KB Updated Sep 17, 2026

File contents

translate — Local Translation via Ollama

Offline-first. Qwen3 via Ollama. No data leaves the machine.

Models (64GB unified memory)

Model	Size	Use
`qwen3:32b`	20GB	Default — best quality
`qwen3:30b`	19GB MoE	Fast — near-same quality
`qwen3:14b`	9.3GB	Acceptable quality, fastest

ollama pull qwen3:32b   # one-time

Run the script

# Translate text to a target language
nu script/translate.nu "Japanese" "technical-formal" "Your text here"

# From a file
nu script/translate.nu "Japanese" "technical-formal" \
  --file /path/to/input.txt

# Round-trip validation (EN -> JP -> EN)
nu script/translate.nu "Japanese" "technical-formal" \
  "Your text" --roundtrip

Workflow checklist

Translation Progress:
- [ ] 1. Choose target language and register (see registers table below)
- [ ] 2. Run: nu script/translate.nu "Language" "register" "text"
- [ ] 3. Verify: does output read naturally? Check for register flattening.
- [ ] 4. For critical text: run --roundtrip and compare back-translation
- [ ] 5. For legal/published copy: use Qwen-MT API instead (see cjk.md)

Registers (Japanese)

Context	Register
Engineering spec, PR	`technical-formal`
Stakeholder report	`formal`
Slack / team chat	`casual`
Customer-facing docs	`polite`
Legal / contract	`keigo`

Thinking mode

/no_think — default, faster, use for technical docs and UI strings
/think — use for literary passages, ambiguous register, irony

CJK detail

For CJK-specific system prompts, register rules, honorific mapping, chengyu handling, and Korean verb ending nuances: see cjk.md

Shell escaping rule

Never interpolate text into jq strings. Always --arg + pipe via @-. The script handles this correctly. See cjk.md for the pattern.

Limitations

Qwen3:32b local vs Qwen-MT API: local strong, Qwen-MT better for legal/published
Proper nouns: verify katakana readings for obscure new names
No pitch accent reconstruction for TTS output

---name: translatedescription: Local multilingual translation via Ollama — 119 languages, no API key, no network, no data leaves the machine. Use when translating to/from Japanese, Chinese, Korean, French, Spanish, German, Portuguese, Russian, Arabic, or any other language. Also for localization, i18n text, and roundtrip validation. Qwen3:32b on Apple Silicon — faster and cheaper than cloud APIs.---# translate — Local Translation via OllamaOffline-first. Qwen3 via Ollama. No data leaves the machine.## Models (64GB unified memory)| Model | Size | Use ||-------|------|-----|| `qwen3:32b` | 20GB | Default — best quality || `qwen3:30b` | 19GB MoE | Fast — near-same quality || `qwen3:14b` | 9.3GB | Acceptable quality, fastest |```bashollama pull qwen3:32b # one-time```## Run the script```bash# Translate text to a target languagenu script/translate.nu "Japanese" "technical-formal" "Your text here"# From a filenu script/translate.nu "Japanese" "technical-formal" \ --file /path/to/input.txt# Round-trip validation (EN -> JP -> EN)nu script/translate.nu "Japanese" "technical-formal" \ "Your text" --roundtrip```## Workflow checklist```Translation Progress:- [ ] 1. Choose target language and register (see registers table below)- [ ] 2. Run: nu script/translate.nu "Language" "register" "text"- [ ] 3. Verify: does output read naturally? Check for register flattening.- [ ] 4. For critical text: run --roundtrip and compare back-translation- [ ] 5. For legal/published copy: use Qwen-MT API instead (see cjk.md)```## Registers (Japanese)| Context | Register ||---------|----------|| Engineering spec, PR | `technical-formal` || Stakeholder report | `formal` || Slack / team chat | `casual` || Customer-facing docs | `polite` || Legal / contract | `keigo` |## Thinking mode- `/no_think` — default, faster, use for technical docs and UI strings- `/think` — use for literary passages, ambiguous register, irony## CJK detailFor CJK-specific system prompts, register rules, honorific mapping,chengyu handling, and Korean verb ending nuances: see [cjk.md](cjk.md)## Shell escaping rule**Never** interpolate text into jq strings. Always `--arg` + pipe via `@-`.The script handles this correctly. See [cjk.md](cjk.md) for the pattern.## Limitations- Qwen3:32b local vs Qwen-MT API: local strong, Qwen-MT better for legal/published- Proper nouns: verify katakana readings for obscure new names- No pitch accent reconstruction for TTS output

Frequently asked questions