Lokalog

🌟 What’s new?

We’ve improved the “Updated keys” filter logic in Workflows. From now on, workflows will only trigger when the base language value of a key is updated. Changes to metadata—like tags, comments, or other non-content attributes—will no longer trigger a workflow run.

🎯 Why it matters

This change helps reduce unnecessary workflow executions, keeping your localization process more focused and efficient. Teams will benefit from fewer distractions and better resource management, especially when working on large projects or with automation-heavy setups. Your workflows will now respond only to meaningful content changes.

🌟 What’s new?

We’re introducing AI-driven scoring for AI translations combined with workflow automation based on quality thresholds (currently in Beta):

• Every AI-generated string in a translation task is now assigned an quality score based on the MQM framework.
• You can set a threshold in your workflow so that only translations scoring below the threshold are sent for human review.
• Manual scoring is also available directly in the editor

🎯 Why it matters

This feature empowers teams to:

• Cut review costs by automating quality filtering
• Make smarter decisions about which content truly needs human review
• Unlock low-priority languages that were previously cost-prohibitive
• Support reviewers with clearer scoring guidance and issue suggestions

👉 Interested in trying this powerful feature early? Join our waitlist for early access to the Beta program →

Want to see it in action? Watch this quick demo where Luke Vella, Staff Product Manager at Lokalise, shows the power of AI Translation Scoring combined with Workflows.

🌟 What’s new?

Teams on Enterprise and Trial plans can now build fully customizable workflows—either from scratch or by modifying existing templates. You can add, remove, reorder, or duplicate steps to design workflows that perfectly match your localization process, with support for up to 15 steps.
Previously, workflows were limited to predefined templates. This update unlocks full flexibility, empowering teams to align automation with their unique needs.

🎯 Why it matters

Custom workflows let you tailor Lokalise to your existing processes, improve automation, and streamline collaboration across teams. Whether you’re scaling operations or refining a complex localization flow, you now have the tools to build workflows that fit how your team actually works.

🌟 What’s new?

You can now create and run workflows on any branch within your Lokalise project—not just the master branch. This update enables teams that work primarily in non-master branches to fully automate their localization process without impacting the integrity of their master branch.

Previously, workflows were limited to the master branch, which restricted adoption for teams using branching strategies to manage content updates. With this change, you can maintain your workflow practices while aligning with your development process.

🎯 Why it matters

Whether you’re managing translations in feature branches, staging environments, or production-ready branches, you can now streamline localization with automated workflows across your entire project structure. This flexibility removes previous blockers and ensures consistent localization workflows wherever your team works.

🌟 What’s new?

Enabled full lifecycle management of AI Tasks through our public API, ensuring parity with all UI-driven capabilities. Customers are now able to create, retrieve, update, and delete AI Tasks programmatically.

🎯 Why it matters

This unblocks mature clients to use automated way of creating tasks and staying informed about their status and progress. In a scope of the update we have confirmed Webhooks app is up to date with progress of AI task events.

🌟 What’s new?

We’ve launched a Beta Program that allows customers to test Custom-trained AI models tailored to their brand voice and content. These models, based on RAG and Fine-Tuned GPT-4o, deliver launch-ready translations with over 90% accuracy on the first attempt by capturing your company’s specific tone and style.

Alongside Custom LLMs, the Beta includes access to AI-powered Translation Quality Scoring, helping you easily assess which translations meet quality standards and which require review—automatically integrated into your workflows.

🎯 Why it matters

This is your opportunity to gain early access to cutting-edge AI localization tools that improve translation quality, save time, and scale efficiently. Joining the Beta helps your team stay ahead by streamlining localization with smarter, brand-aligned translations.

🌟 What’s new?

It now enables users to easily plug in custom pre-processing logic (for file uploads).

🎯 Why it matters

You can fine-tune how translation files are handled before they’re uploaded to Lokalise. Strip metadata, normalize formatting, or apply project-specific rules.

🌟 What’s new?

Now users can customize Git-related settings easily.

🎯 Why it matters

With this improvement you get more control over how Lokalise interacts with your repositories, adjust commit messages, and use custom tokens.

🌟 What’s new?

We’ve significantly enhanced the user experience for creating tasks, especially in large projects.

Previously, users experienced delays (or even freezes) after submitting a task form.

Now, the task creation process is much faster and smoother, with the UI updating immediately.

In internal tests, we observed up to a 2300% improvement in UI reaction time, and issues with the interface getting stuck are now resolved.

🎯 Why it matters

This update greatly enhances the manual task creation experience, particularly for teams managing large projects or wide scopes. Users can now work more efficiently without frustrating delays.

🌟 What’s new?

Ever wondered what your workflow has been up to? Say no more, your new status reporting tab in Workflows is here! Filter by date, status, language, or title and track everything that happened. No more guessing. Just clarity.

🎯 Why it matters

It’s all about trust. Status reporting gives you full visibility into what’s running and when—so you can track progress, and stay in control of your automation.