What Makes Raiverb a Unique Translation & Localization Tool

Raiverb is accessible to anyone and easy to set up. It shares common features with classic CAT tools, so any experience with these tools will help, but is not mandatory. Just as with anything AI, the output depends on the way the job is setup, so reading the manual will help.

Raiverb is designed to work with high volumes. Since the software runs on your computer, the limit is mostly your CPU and memory. There’s a hard cap of 8000 Translation Units (Translation Units can be understood as paragraphs). We believe hardly any job should require more than this, but if we ever get feedback saying the opposite, we’ll publish the workaround.

Translation Memories and Glossaries with tens of thousands of entries work without issue. Again, you’ll have to make sure your machine can manage it, test tools are available.

You can currently import: .XLSX, .DOCX, .CSV, .PDF, .ASS, .SRT, .JSON, .TMX, .XLIFF, .POT formats. You can also import XML-based formats and let Raiverb try to find the importable content.

PDF documents are still considered experimental, some more work will be needed before PDF can be saved convincingly into their original format.

You can also import pictures, as Raiverb contains an Optical Recognition feature that will try to extract text from most common formats, and also from the Internet if you input a URL. Work is being done to import all sorts of additional formats in the future.

Translations require an internet connection to run jobs. However, some features can run offline, including the images importer and the Translation Memory converter (under the Utilities tab).

Raiverb is free to download and use. You also do not need an account to do either.

However, note that the main feature of Raiverb is automated translation/proofreading, and both of these functions require an account.

A free trial server is available to test the paid functions.

Here’s what you can do for free in Raiverb:
– Extract text from pictures, including from the Internet.
– Convert your content into CAT-ready Translation Memories. Append content to Translation Memories.
– Convert Chinese from Simplified and Traditional and vice-versa.
– Use as a simplified CAT tool.

These cool features are completely free and do not require an internet connection.

One last thing: whenever you chose to support us and use our paid translation access, either through subscriptions or direct tokens purchase, all the records you make are stored and accessible offline, stored on your computer, and accessible regardless of your future status and credits.

We chose to keep our features out of arbitrary paywalls.

You may want to read this: https://www.bouncerock.com/raiverb/knowing-the-system/

It is up to you to decide if you need a human translator. However, Raiverb was not built from the idea of replacing humans, but facilitating the workflow for humans and optimizing the output of AI for translation.

Raiverb was designed around the concept that context and consistency is essential. This is the same concept as every other CAT software around, and this is also the very reason why generative models or any other type of translation machine as we know it now should not replace human input.

One last thing: even human translators need human proofreaders. Raiverb has the option to proofread translations that works well, but ultimately, we’d still advise you to use it to facilitate the proofreader’s job, not to replace it.

The main difference between Raiverb and Machine Translation tools, is that Raiverb can take advantage of the context the user chose to input into it, and is able to maximize consistency and accuracy through using Translation Memories and Glossaries.

Hard work is being put in to make Raiverb work seamlessly efficiently with the largest range of jobs and industries possible. Raiverb was designed with the video game industry in mind first, but works great with technical projects. It can also produce satisfying content with more creative jobs, but it will all ultimately boil down to the way you set it up. It is strongly advised to read the manual before starting any serious job.

We are still in beta, so we are not charging yet. Eventually, words will be charged per volume fed to the server. This includes the source text, but also all the other data, including String IDs, alternate languages, extra requests and the likes. This also includes the translation and the comments generated. The software will provide you with a guesstimate of the expected cost before you run a job.

There are no discounts for repetitions, except for perfect matches with either the Translation Memory or the Glossary. Meaning you don’t get charged twice for the same content even if your project has a lot of repeated content.

contact@bouncerock.com