Introduction
Co-op Translator began as a simple CLI tool for documentation translation. Over the past year, it has evolved into a GitHub Actions–powered workflow that helps automate translation across Microsoft’s open-source learning repositories.
What started as a lightweight translator is now a widely used automation layer that keeps multilingual content continuously in sync, making Microsoft’s educational resources accessible to global communities without creating extra work for maintainers.
Today, we’re excited to announce the release of v0.10.0, which introduces Jupyter Notebook support, expanded language coverage, and a more reliable translation engine.
Status and Utilization
Adoption at Scale
- 27,700+ downloads from PyPI as of August 2025
- Actively used in multiple Microsoft OSS repositories
Community Feedback (June–August 2025)
- 37 reports received on translation quality
- 35 resolved, with 2 pending (Mermaid diagrams and <details> tag translation)
Supporting Microsoft Learning Content
With GitHub Actions integration, Co-op Translator helps keep translations in sync automatically.
Examples include:
Global Reach
Supports 46 languages, including French, Chinese, Japanese, Korean, Spanish, Tagalog, Burmese, and more — enabling learners worldwide to access resources in their native language.
Workflow Impact
Multilingual documentation often lags behind source updates, leaving learners with outdated instructions or broken links.
Co-op Translator addresses this by automating updates at commit-time, ensuring translations remain aligned with source content across every release cycle.
Community Activity (June–August 2025)
✨ New Features & Updates
- New language support: Added Ukrainian, Burmese, and Lithuanian, each contributed by native speakers to ensure real-world accuracy
- Notebook support: .ipynb translation with hash-based change detection, keeping notebooks in sync while preserving markdown, code, and outputs
- Smarter evaluation (Beta): Experimental GPT-based anomaly detection flags suspicious translations to reduce manual review workload
- Engine refinements:
- Markdown translator upgraded with dynamic token-based chunking → more consistent results in dense or link-heavy docs
- Image translation pipeline enhanced with structured OCR mapping, preserving layout and readability
Community Highlights
Progress was made possible with contributions from the Microsoft community:
- Minseok Song (Maintainer, Microsoft AI MVP) — translation engine improvements (dynamic chunking), Jupyter Notebook hash detection, structured image output
- Vladislav Antonyuk (Microsoft MVP) — Ukrainian language support
- Hiroshi Yoshioka (Microsoft MVP) — Japanese translation quality improvements
- Will Huang (Microsoft Regional Director / Microsoft MVP) — Jupyter Notebook translation support
Alongside these, other contributors refined error handling, expanded language coverage, and improved documentation consistency.
Conclusion
Co-op Translator continues to evolve with the community.
If you spot awkward or inaccurate translations, please share feedback via GitHub issues so we can keep improving together.
With the release of v0.10.0, projects can now confidently adopt it across docs, notebooks, and images.
