Goman.live - A Service for Localizations and Translations with AI
By Admin
Goman.live - A Service for Localizations and Translations with AI
Greetings, everyone! Today, I will share more details about the service I'm creating, Goman.live - what it does, how it works, who it might be useful for, and where we are heading in the future. ---
What is Goman.live?
First and foremost, Goman.live is a service for managing localizations and translations using modern artificial intelligence models. It helps support multilingual web and mobile applications, games, or simply translates texts faster and without unnecessary routine. The service operates in two main modes:
- through MCP (Model Context Protocol), if you want to do everything through agents in Copilot or Cursor,
- and through its own editor, where you can manually manage localizations, check translations, and add context and rules. ---
Who is it suitable for?
Goman.live is suitable for those who work with content and want to make translation more intelligent and organized:
- 👨💻 Developers of multilingual applications.
- 👩💼 Teams working through GitHub Copilot, Cursor, or other MCP clients.
- 🗣️ Translators and content managers who need to quickly update translations.
- 🌐 Anyone who wants to have a "smart" localization editor at hand. ## About MCP
Managing localizations through MCP allows you to save, update, and delete keys and localizations through agents. The agents know which languages your application supports and immediately create translations for all these languages.
The service checks whether a record exists before creating or changing it, which helps avoid duplicates and collisions. This approach greatly simplifies content management and allows for quick updates of texts "on the fly." The only thing that still requires attention is the excessive generation of keys and translations that are not used. We are working on automatically checking and removing such localizations. However, even now, the MCP approach makes localization management fast and transparent. ---
About the Editor
The editor is the second most important tool in Goman.live. It not only allows you to create new localizations but also to edit existing ones, make translations into multiple languages at once, find errors, work with context, and much more. With the editor, you can translate even large texts, taking into account context, checking for errors, and saving results. It is suitable for both technical and ordinary texts.
Supported Languages
There is no complete list of languages in the service, and that's a good thing. We use models capable of working with dozens and hundreds of languages, so there are practically no limitations. The quality of the translation depends on the text itself, the prevalence of the language, and whether there are similar translations in your database.
Translations through the editor often turn out better than through ordinary tools because they take into account context and previous results. And if you work through MCP, the quality depends on the model your agent uses, such as Claude, GPT, or others. 
About Context and Prompts
Context is one of the most important aspects of translation. In Goman.live, you can use both previous translations (RAG) and your own prompts, where you set the rules and limitations yourself. Such a prompt can replace a "glossary"—it helps the model understand your style, terminology, and even tone. Prompts support versioning, allowing you to save different versions and revert to older ones.
Additionally, you can create small context extensions directly in the editor, for example, when you need to clarify the translation of a particular word or phrase. Let's talk about error correction. Goman.live offers two simple yet useful features for working with errors: You can enable automatic checking during translation (checkbox "Use Syntax Check") – after translation, the model checks the text and makes corrections independently. 2. Or perform a manual check in the editor – all suspicious spots are highlighted, and you can decide whether to accept the suggested changes yourself.
This way, you always keep control in your hands.
Limitations and Features
Since the service operates on LLM, it has its own features. Firstly, it is the context size: the model can only process a certain amount of information at a time (on average about 8-10 thousand tokens). If the text is too large, it's better to divide it into logical parts so that the model can better understand the structure and context, resulting in more accurate translations.
When using RAG, the context may also be incomplete because it stores individual pieces of text, and the search occurs through them. Therefore, it's better to translate in smaller blocks and avoid overloading the model.
This is not a disadvantage, but rather a characteristic of how LLMs work—they just need to "understand" the text within their context window. However, this allows for stable results and easy processing of translations.
Additional Functionality
Goman.live is not just about localization. The service allows testing prompts, viewing results, saving them in different versions, and even working in a team. This is more thoroughly described in the service documentation.
Storing Prompts and Versioning
You can store up to four versions of a single prompt and switch between them with a single click. This is very convenient if you want to experiment with formulations or translation rules. Prompt versioning, when used properly, can significantly reduce token consumption. It can also be useful for testing and finding the most suitable prompts. We have added a prompt testing system where you can immediately run execution requests and get results.
Collaboration
If you are not working alone, simply add your colleague's email. He will gain access to your project, and you will be able to work together on localizations and translations. The limitations depend on the tariff plan. Tokens for translation are taken from the owner's account - that is, the one who "shared."
About Tokens
When translating through the editor, LLMs are used, so tokens are counted - both input and output. Usage results are calculated during requests and updated after a short time - usually one or two minutes. This is necessary for transparent pricing. If you work through MCP, tokens are not used at all because everything is translated through an agent (Copilot, Cursor, etc.).
How to Save Tokens
First of all, it should be noted that functions such as error checking and correction require additional tokens. ### Tokens Can Be Lost Tokens can also be lost when using overly large prompts, as they get fed into the model. Therefore, it is sometimes better to manipulate multiple prompts or have several different versions rather than one large one.
How to Get a Discount
We have a referral program: bring a new user, and you will receive up to a 90% discount for one month. It's a small perk, but it's nice 🙂. So, if you're interested, write to us at email. ---
About Data Preservation
All your data is stored on our own servers. We perform a backup every six hours and do not share information with third parties, except for the models (LLM) that process translations. Data may still be retained for some time after the application is deleted—this is done to allow for the possibility of restoration upon personal request. Additionally, you can export localizations at any time in the format provided by the service. Data can be uploaded through MCP via a form (one record at a time) or in bulk from a file in the required format.
About Cooperation
We would like to emphasize that we are open to new ideas and collaborations. If you have a service or business that can be integrated with Goman.live, let us know! Or if you're simply interested in something, we are ready to create custom plans, API integrations, and even separate solutions tailored to your needs. ---
What's Next
We are already working on the following:
- The ability to use prompts directly in GitHub Copilot and Cursor.
- Choosing the translation model and using custom API keys.
- Improving RAG and automatic localization updates.
- And many more ideas that will make working with translations even more convenient.
PS
Thank you very much for reading! I would be very grateful for comments, feedback, and suggestions. Your feedback helps the service grow and improve.
👉 Visit goman.live to see how everything works in practice 🌍