Tool Ren'Py RenLocalizer - Automatic Translation Tool for Ren'Py Games

[csmail12]

I’m really sorry to hear about everything you’re going through. You mentioned your situation before, and I didn't want to press on it too much because it’s such a heavy topic, but I felt the need to reach out. It’s heartbreaking that you have to deal with power cuts and bombings while just trying to work on your projects. No one should have to go through that.

Regarding the program, I wish I could give you a 100% guarantee for the API engines right now. But just like you, I have my own constraints. Since I’m developing this on a tight budget, I have to prioritize my basic needs over buying API credits for extensive testing. I’m basically trying to fix things blindly with whatever free tier I can find.

I really appreciate you sticking with my tool despite the hardware limits and the situation in your country. I’ll keep doing my best to optimize it, even with my limited resources. Stay safe out there—I truly hope things get better for you soon.

I can see that you’re working on the project conscientiously and from the heart. I’d love to support you with a larger amount, at least for the subscriptions, but unfortunately, I can’t even afford my own subscription right now.


That said, I looked into the DeepL subscriptions. I thought I could at least support you that way. But now I see that what you need costs 50 euros… I didn’t realize it would be that ridiculously expensive.

 

[LordGamer00]

I can see that you’re working on the project conscientiously and from the heart. I’d love to support you with a larger amount, at least for the subscriptions, but unfortunately, I can’t even afford my own subscription right now.


That said, I looked into the DeepL subscriptions. I thought I could at least support you that way. But now I see that what you need costs 50 euros… I didn’t realize it would be that ridiculously expensive.

You’re absolutely right about DeepL. It really is the 'sweet spot' for game translations—fast, reliable, and way less complex than dealing with LLM prompts. I’ve realized that focusing too much on AI might be distracting me from perfecting what most people actually use.

And yeah, that 50 Euro price tag for full API access is exactly why I’ve been struggling! It’s a pretty steep cost, especially for an open-source project where everything is shared for free. Paying that much out of pocket just to test features is a bit too 'salty' for a solo dev. That’s why I’ve been trying to 'hack' my way through with free limits, but it definitely makes thorough testing a nightmare.

I really appreciate the thought of supporting me, but I want to be clear: I don't want anyone in a tight financial spot to feel like they need to donate. I know exactly what it’s like to struggle financially. I only added the donation part for those who truly have the extra means. To be honest, it’s not going great on that front, but that’s okay with me.

This project actually started with a very simple goal: I just wanted to build a Ren'py translation tool for myself. I decided to share it once it was ready, but I probably would have abandoned it a long time ago if it weren't for users like you. Your feedback is what brought the project to this point and what keeps me going.

So, I’m going to take your advice and double down on DeepL to make it as rock-solid as possible. Thanks for keeping me motivated!

 

[denadm]

And here's another thing I remembered: you need to add more languages to the input language, because the “auto” option doesn't work at all.
I tried Spanish to Russian in auto mode, and it's nonsense. At least add the main European languages. Less common than English are Spanish and Italian, and even less common is German. I've hardly ever encountered any others.

 

[LordGamer00]

And here's another thing I remembered: you need to add more languages to the input language, because the “auto” option doesn't work at all.
I tried Spanish to Russian in auto mode, and it's nonsense. At least add the main European languages. Less common than English are Spanish and Italian, and even less common is German. I've hardly ever encountered any others.

I hadn't focused on that part at all because I thought the automatic option would suffice, but since it can be nonsensical, I'll add it, of course.

 

[denadm]

I recommend checking your parser. It scans a lot of unnecessary stuff, such as folders: cache, renpy, whose contents do not need to be translated at all. Or maybe you should give the user a choice of folders to parse. I would prefer only files rpy(rpyc) in the “game” folder to be parsed excluding unnecessary system folders and files *.rpyb.

After parsing, I found about 10,000 lines of unnecessary translation in the translation file that are not related to the plot of the novel.

 

[LordGamer00]

I recommend checking your parser. It scans a lot of unnecessary stuff, such as folders: cache, renpy, whose contents do not need to be translated at all. Or maybe you should give the user a choice of folders to parse. I would prefer only files rpy(rpyc) in the “game” folder to be parsed excluding unnecessary system folders and files *.rpyb.

After parsing, I found about 10,000 lines of unnecessary translation in the translation file that are not related to the plot of the novel.

Thank you for your valuable feedback. You are absolutely right; the parser is currently scanning some files and folders that should be excluded from the translation process.

To be honest, this issue is a leftover from the period when I was focusing heavily on the Google Translate engine integration. Since it didn’t cause any critical errors within the game, I didn't prioritize fine-tuning the filters at that time. However, now that we are integrating various other translation engines, filtering these unnecessary files has become a priority to avoid wasting API limits and reducing costs.

Regarding the manual folder selection, I’ve been trying to stick to a "select, translate, and play" philosophy to keep the user experience as simple as possible. That is why I haven't given the user full control over the parsing paths yet. However, I will completely overhaul the parsing system to ensure that system folders (like cache and renpy) and unnecessary file types (like *.rpyb) are automatically excluded.

I appreciate your contribution to making the program more efficient!

 

[LordGamer00]

Hello everyone,

I have been working on the RenLocalizer project for some time now and have reached the end of version 2.5.0. I have largely completed the technical preparations and am currently in the final testing phase to ensure the version runs as smoothly as possible. Barring any issues, I will share the update with you as soon as the tests are complete.

What has changed in this version in general?

• Smart Skip & Resume: I have attempted to optimise the translation process a little further. The system can now recognise and skip previously translated lines and resume from where it left off in the event of an interruption. This aims to save both time and quota.
• Parser Update (Filtering): I have overhauled the filtering system to reduce the accumulation of unnecessary technical lines in the translation files. It now largely excludes system folders such as cache and renpy, as well as certain code files that should not be translated.
• Error Handling (Retry): I added a simple retry mechanism to address situations where some AI models return the text verbatim instead of translating it.
• Language Support: I have expanded both the source and target language lists; approximately 90 languages are now supported.
• Local Models: I have added Jan.ai support to the Local LLM side and several categories to simplify model selection.
• Engine Optimizations: Some optimizations were made for translation engines.

A Quick Note: As you know, I am actively continuing to work on the project. Although I test it in various scenarios, there may be overlooked errors or unexpected situations in different game engine versions. If you share any issues you encounter or ideas you have about how things could be improved, I can focus on them in the next updates.

Thank you for your support and patience! ❤

 

[NaveenKrishnanC]

That’s why I’ve been trying to 'hack' my way through with free limits, but it definitely makes thorough testing a nightmare.

You can check your chrome or edge browser for more stable and better free translation api.
Though it is indeed hacking as abusing free api.

 

[LordGamer00]

You can check your chrome or edge browser for more stable and better free translation api.
Though it is indeed hacking as abusing free api.

Thank you for the suggestion! You are absolutely right, and I appreciate your input.

Actually, RenLocalizer is already using a similar approach for Google Translate. We are directly utilizing Google's public web endpoints (the same ones used by browsers) to provide free translation without requiring an API key.

However, applying this same method to other engines involves different challenges:

1. LLM Architectures (AI): Modern models like Gemini or GPT function very differently from traditional web translators. They require strict session handling and authentication (tokens/cookies) that cannot be easily simulated via simple web requests. Trying to bypass their official APIs creates a very unstable experience that breaks frequently.

2. Strict Security (DeepL etc.): While Google's public endpoints are relatively lenient, services like DeepL are extremely aggressive against automated access. They employ strict IP bans and CAPTCHAs that trigger almost immediately when accessed unofficially. Without a complex proxy network, user IPs would get banned very quickly.

That's why for version 2.5.0, I prioritized the "Smart Skip" and "Resume" systems. Instead of relying on fragile methods to get "free" access for every engine, these features optimize the process so that official APIs and Local LLMs are only used for new content. This minimizes costs and maximizes stability.

Thanks again for sharing your thoughts!