We are excited to announce that Gaia 3.0 is complete, and it represents the system we set out to build from the beginning. When we first introduced Gaia in April, our focus was clear: to create the most flexible and powerful translation management system for game localization. In May, with Gaia 2.0, we expanded that foundation with Adventure and Forge, bringing content creation directly into the localization workflow.
From translation management to engine integration
With Gaia 3.0, we are taking another major step forward. Gaia now offers native plug-in integrations with two of the world’s most important game engines: Unreal Engine and Unity. This allows both engines to read content directly from Gaia’s translation memory with speed, reliability, and efficiency.
This changes how localization can be connected to game production. Instead of treating localization data as an export that has to be manually prepared, transferred, validated, and reworked, Gaia 3.0 makes the translation memory a live operational source for the game engine itself.
Designed for localization engineers
The integration process was designed specifically with localization engineers in mind. Connecting Gaia to Unreal Engine or Unity is straightforward, predictable, and built around the realities of localization implementation work.
Game developers should be able to focus their time on gameplay, performance, systems, and player experience. Localization teams should have the tools they need to own localization integration with confidence. Gaia 3.0 was built around that principle: localization should empower development, not slow it down.
Local LLMs, inside your own environment
The second major announcement in Gaia 3.0 is local LLM support. Gaia can now connect to local large language models, opening a new path for AI-assisted translation, review, and content workflows without requiring teams to send sensitive production data to a cloud AI provider.
This is a breakthrough for power, flexibility, quality, and, most importantly, privacy. Local LLMs give teams more control over how AI is used, where content is processed, and how localization data is protected. For studios working on unreleased games or sensitive IP, that control matters.
- Quality: Modern LLMs can deliver highly contextual translation and language support, especially for dialogue, narrative text, UI copy, and content where tone and intent matter. Gaia keeps human review at the center, while giving teams access to AI output that can be more flexible and context-aware than traditional machine translation.
- Privacy: Local LLMs run on hardware your organization controls and can be configured for offline use. When Gaia is deployed locally and connected to a local model, unreleased game content can remain inside your own environment, instead of being sent to external AI services.
- Value: Many powerful local models are available without monthly subscription fees. Once the model is downloaded and Gaia is connected to it, teams can use AI-assisted workflows without adding per-request cloud AI costs. This is aligned with Gaia’s philosophy: reduce recurring dependencies, increase ownership, and give organizations more control when they choose to purchase the source code of the system.
AI-assisted workflows, with humans in control
At Tomori, we believe AI is a powerful tool, not the whole solution. Gaia was built to help humans do their best possible work and, when they choose to, use AI-assisted workflows to improve the final result with less effort.
Local LLM support makes that philosophy stronger. Teams can use AI where it adds real value, keep editorial judgment in human hands, and preserve control over how sensitive content is processed.
Rewrite translations directly in the grid
In the translation grid, teams can quickly rewrite translations to make them shorter, more formal, or better adapted for UI. The process is straightforward and fully integrated into the translation workspace, so linguists can improve text without leaving the segment they are reviewing.
Draft quests in Adventure
In Adventure, Gaia’s content creation tool, teams can draft entire quests with the help of AI, then adapt the content manually to match the game’s tone, structure, and production needs. Once the content is ready, it can move into the translation workflow with the push of a button.
Translate full project content from the product card
Gaia can also use LLMs to translate full project content directly from the product card. The same workflow that previously supported OpenAI, Google Translate, and other MT or AI providers now supports local LLMs as well, giving teams another way to combine automation, review, and ownership.
Conclusion
Gaia started as a translation management system. With Gaia 2.0, it became a content creation and localization production environment. With Gaia 3.0, it now reaches directly into the engines where games are built and into the local AI environments teams choose to control themselves.
That is the direction we believe localization software must take: closer to production, closer to engineering, closer to creative systems, and more respectful of the ownership teams need over their own content.
On top of the features described above, Gaia 3.0 also has over 30 new features. They are all available now, and you can see them in the Complete Gaia Gallery.
Related tutorials
You can read the tutorials for integrating Gaia with Unity, Unreal Engine, and local LLMs below.