Recipe: Developer Artifacts

yaku translates the human-readable parts of developer output while preserving timestamps, error codes, file paths, and technical identifiers.

Error messages

echo "Fehler: Verbindung zur Datenbank auf localhost:5432 verweigert" | yaku --to en
# Error: connection refused to database at localhost:5432

Log files

Pipe logs through yaku to understand them in English:

# Translate recent logs
tail -50 /var/log/app.log | yaku --to en

# Translate error lines only
grep "ERROR" app.log | yaku --to en

Stack traces

cat error.log | yaku --to en --context "Python stack trace"

yaku preserves file paths, line numbers, and function names. The error message and description text are translated.

API error responses

curl -s https://api.example.com/users/999 | yaku --to en --format json

If the response is JSON, use --format json to translate only the string values (error messages) while preserving status codes, keys, and structure.

Docker and Kubernetes output

# Translate pod status descriptions
kubectl describe pod my-pod | yaku --to en --context "Kubernetes"

# Translate Docker build errors
docker build . 2>&1 | yaku --to en --context "Docker build output"

Tips

• Add --context for domain accuracy. "Python error log", "Kubernetes events", or "nginx access log" helps the LLM preserve the right technical terms.
• Use --verbose to check token usage. Large log files consume more tokens. Consider filtering with grep or tail first.
• Don’t translate logs you need to search later. Translated logs are useful for understanding, but keep the originals for grep-based debugging.