I need a translation tool that runs locally and does not need an internet connection.

Running translation completely locally without an internet connection is very practical. Depending on whether you want a traditional lightweight translation engine or a modern AI-driven language model, here are the best tools available:

1. Argos Translate (Best Lightweight / Traditional Engine)

Argos Translate is an open-source, offline neural machine translation library written in Python. It works entirely on your local machine using OpenNMT and doesn't ping any external servers.

• How it works: You download specific language package files (.argosmodel ) for the language pairs you need. It can even automatically pivot through intermediate languages if a direct pair isn't installed.
• Interface: Can be used as a Python library, via the command-line, or through community-made GUI/desktop apps.
• LibreTranslate: If you want a self-hosted local web server or API endpoint that runs offline, you can look at LibreTranslate , which is built directly on top of Argos Translate.

2. TranslateGemma / Local LLMs via LM Studio or Ollama (Best for Context & Nuance)

If you need high-context localization, tone adjustments, or support across dozens of languages simultaneously, running an open-source translation-focused LLM locally is a great choice. Google's TranslateGemma series is explicitly optimized for offline, high-accuracy multi-language translation.

• The Sizes: The smaller 4B parameter version works on mobile/lightweight hardware, while the 12B version runs smoothly on standard laptops.
• How to run it: Download an application like LM Studio (or use Ollama ), search the model hub for "TranslateGemma" or relevant translation models, download the GGUF file, and run it locally with zero internet required.

3. Mobile Alternatives (If you meant mobile/tablet)

If you are looking to do this on a phone or tablet rather than a desktop:

• Google Translate / Microsoft Translator: Both official mobile apps allow you to download full offline language packs for on-the-go translation without data or Wi-Fi.

If you can tell me what operating system you're using (Windows, macOS, Linux) and whether you prefer a desktop app, a command-line tool, or a Python script , I can give you exact installation steps.

Yes. If you want translation that runs entirely on your computer after the language models are installed, my top recommendation is Argos Translate.

Best options

• Argos Translate — Best overall for a simple local/offline setup. It is open-source, supports a GUI, command line, and Python, and keeps translation on-device once the language models are installed. It supports languages including English, Spanish, French, German, Chinese, Japanese, Korean, Arabic, Hindi, Portuguese, Russian, Ukrainian, and others.
• LibreTranslate — Best if you want a local web interface or API that other programs can call. It is built on Argos Translate and can operate offline after installation/models are downloaded.

What I'd choose

If you're looking for something like Google Translate, but with no internet connection and no text leaving your computer, start with Argos Translate. You download the required language models while online, then you can disconnect and translate locally.

If you tell me Windows, Mac, or Linux and which languages you need (e.g. English ↔ Spanish), I can give you the simplest installation instructions.

If you need a translator that runs entirely on your own device after installation (no text sent to the cloud), these are good options:

My recommendation by use case

• Simple text translation on a laptop: Argos Translate
• A private Google-Translate-like service on your network: LibreTranslate
• Lots of languages, including less common ones: look at NLLB-based or other multilingual local models (hardware requirements are higher).
• Professional documents where nuance matters: a local LLM translator may produce more natural wording, though it requires a stronger computer.

A typical setup is:

1. Install the translator while online.
2. Download the language models you need.
3. Disconnect from the internet.
4. Translate locally with no data leaving the machine.

If you tell me your device (Windows/macOS/Linux/Android/iPhone), languages, and whether you need voice or just text, I can narrow it down to the best choice.

Running a translation tool completely offline, locally on your machine, is a great way to ensure 100% privacy and zero data leakage . Depending on whether you prefer a ready-to-use desktop app , a local web interface/API , or a command-line tool , several excellent open-source options fit the bill.

• LibreTranslate — Best for a local web UI & API.
  • What it is: A free, self-hosted open-source machine translation service powered by offline neural machine translation packages. Once you download the language models, it spins up a clean web interface accessible via your browser at localhost:5000.
  • How to run it: You can install it via Python (pip install libretranslate ) or run it via Docker. You can even restrict it to load only the languages you care about to save RAM.
• Argos Translate — Best for a lightweight Python library or GUI.
  • What it is: The underlying open-source translation engine that LibreTranslate is built upon. It is an offline neural machine translation library written in Python that utilizes OpenNMT. It handles local .argosmodel packages and can pivot through intermediate languages automatically if a direct language pair isn't installed.
  • How to run it: Usable as a command-line tool, a Python package, or via community desktop GUI wrappers.
• AI Translate (Windows Store) — Best for a native Windows desktop app.
  • What it is: A native application designed specifically for Windows that utilizes on-device local AI processing to handle fast, secure, and fully offline translations.
  • How to run it: Available as a direct install from the Microsoft Store, requiring standard runtime prerequisites (.NET Desktop and VC++ runtimes).
• Local LLMs (via Ollama / LM Studio) — Best for high-accuracy contextual and nuanced translation.
  • What it is: Running a compact open-source model (like Llama 3, Mistral, or Gemma) locally on your hardware. While heavier than traditional machine translation models, modern localized LLMs excel heavily at localization, tone adaptation, and translating complex or colloquial text entirely offline.

To help narrow down the best fit, tell me:

• What operating system are you using (Windows, macOS, Linux)?
• Do you prefer a graphical user interface (GUI) , a command-line tool , or a local web page?

Yes. If you need translation that continues to work with Wi-Fi completely disabled, I’d look at these:

1. Argos Translate — best simple choice

• Runs entirely on your computer.
• Open source and free.
• Has a GUI, command-line interface, and Python API.
• You download the language models once; afterward, translation can be done offline.
• Supports languages including Spanish, French, German, Chinese, Japanese, Korean, Arabic, Hindi, Portuguese, Russian, Ukrainian, and others.
• Available for Windows, macOS, and Linux.

My recommendation if you just want a local translator without much complexity.

2. LibreTranslate — best if you want a local web/API interface

LibreTranslate is built on Argos Translate and can be self-hosted. It provides a browser-based interface as well as an API, and is explicitly designed to be offline-capable.

It's particularly useful if you want to have a translation service running on your own PC/server and allow other local applications to use it.

3. Bergamot / OPolyglot — worth considering for a desktop translator

Bergamot is the technology behind Mozilla's local translation system. For example, OPolyglot is a newer desktop application using Bergamot that supports fully offline translation and even has a screen-translation/OCR feature.

What I'd choose

One caveat: local models generally aren't as good as the best cloud services such as DeepL or Google Translate, particularly for nuanced writing.

If you tell me Windows, Mac, Linux, Android, or iPhone, and the languages you need, I can recommend the best one and give you the exact offline installation steps.

If you need a translation tool that runs 100% locally and offline on your computer or device, several open-source and dedicated privacy-focused desktop options fit your needs.

• Argos Translate - An open-source offline translation library written in Python. It uses neural machine translation and lets you download specific language package zips to run entirely on your local machine via the command line, a Python script, or community-built GUIs. You can explore the project on the Argos Translate GitHub Repository.
• Crow Translate - A lightweight, cross-platform desktop translator interface that can utilize local offline engines (like Argos or standard translation packages) to perform text, speech, and OCR translation without pinging the cloud.
• translateLocally - A desktop application designed for local machine translation, letting you download specific language pairs so you can translate text privately right on your hardware.
• Local LLMs (Ollama / LM Studio) - If you have a machine with decent RAM or a dedicated GPU, you can run an open-source multilingual model locally (such as Gemma or Qwen) through a tool like Ollama . Once the model weights are downloaded once, you can prompt the local AI to translate text completely offline with high contextual accuracy.
• Lingvanex Desktop / Euroglot NMT - Paid/freemium commercial desktop applications built specifically for Windows and macOS that offer robust offline neural machine translation packages for corporate or privacy-sensitive document translation.

To help narrow down the best choice, let me know:

• What operating system are you using (Windows, macOS, Linux)?
• Do you prefer a graphical user interface (GUI) or a command-line tool?

If you want a translation tool that runs entirely on your own computer with no internet connection, these are some of the best options:

My recommendation

For a normal offline translator, start with Argos Translate:

• ✅ No account
• ✅ No cloud connection after installation
• ✅ Runs on ordinary laptops
• ✅ Supports many language pairs
• ✅ Can translate text from a GUI or command line
  

Typical setup:

1. Install Argos Translate while online.
2. Download the language packs you need (for example English ↔ Spanish).
3. Disconnect from the internet.
4. Translate locally.

If you need higher-quality translations closer to Google Translate/DeepL, you may want to look at running larger local AI models (for example translation models through tools like CTranslate2), but they usually require more RAM/storage.

If you tell me your operating system (Windows/macOS/Linux/Android/iPhone) and the languages you need, I can suggest the best offline setup.

Here are the best local, offline-capable translation tools available for desktop and mobile, ranging from lightweight open-source engines to advanced AI models:

• Argos Translate - An open-source offline translation library written in Python that uses OpenNMT models. It runs entirely on your local machine, supports a wide array of languages, and can be used as a command-line tool, Python library, or via graphical interfaces.
• LibreTranslate - A free, open-source machine translation API that can be completely self-hosted and run offline (often powered underneath by Argos Translate). You can spin it up locally via Docker or Python for a private, web-based translation UI.
• translateLocally - A fast, secure desktop app built for local machine translation. Powered by the Bergamot/Marian neural machine translation engine, it lets you download specific language pairs via the internet once, and then disconnect completely to translate entirely on-device.
• Firefox Translations - Built directly into the Mozilla Firefox browser, this extension processes entire web pages locally on the client side. Once the initial language models are cached, it does not require an active internet connection to translate text.
• RTranslator - An open-source, offline real-time translation app for Android that utilizes Meta's NLLB (No Language Left Behind) model. It is designed for private, face-to-face conversational translation without sending any data to the cloud.

To help narrow down the best fit, let me know:

• What operating system are you using (Windows, macOS, Linux, Android, iOS)?
• Do you need a Graphical User Interface (GUI) or a command-line/API tool?

Running translation entirely offline on your own hardware is a great way to maintain total privacy and independence from cloud APIs. Several powerful, open-source local translation tools and engines fit this description.

• LibreTranslate : An open-source, self-hosted machine translation engine that runs completely offline via Python or Docker. It supports dozens of languages and can be installed on your local machine with a simple command line setup. You can explore the project on the LibreTranslate GitHub Repository.
• Bergamot Translator : A client-side machine translation project funded by the European Union that powers offline translation in browsers and desktop tools like Firefox and translatehtml. It relies on MarianNMT for fast, local CPU-based neural machine translation. More details are available on the Bergamot Project Site.
• Argos Translate : A neural machine translation library written in Python that works offline and uses the packages from LibreTranslate under the hood. It can be integrated into scripts or used via a graphical interface. Learn more via the Argos Translate GitHub Repository.
• Calypso / Ollama-based approaches : You can run open-source large language models (like Llama 3 or Mistral) locally using tools like Ollama and prompt them to translate text offline with high accuracy, though it is more resource-heavy than dedicated translation engines.
• Translation Add-ons / Desktop Apps : If you prefer a ready-to-use desktop application rather than coding or server setups, open-source projects or local frontends for Argos/LibreTranslate provide a seamless GUI experience.

Would you prefer a command-line tool , a desktop application with a graphical interface , or a Python library to integrate into a script?

Yes. If you want translation that stays entirely on your computer and continues working with Wi-Fi/Ethernet disconnected, there are several good options.

My top recommendation: Argos Translate

• Fully offline after you download the language packs.
• Open source and free.
• Works on Windows, macOS, and Linux.
• Has a GUI, command-line interface, and Python API.
• You download only the language pairs you need.
• It can translate through an intermediate language when a direct language pair isn't installed.

The main drawback is translation quality: it's generally less polished than Google Translate or DeepL, particularly for nuanced or specialized writing.

Other good choices

TranslateLocally — worth considering if you want a lightweight desktop translator that emphasizes speed and CPU-only operation. It runs locally on Windows, macOS, and Linux.

LibreTranslate — particularly good if you want a local translation server/API rather than just a desktop application. It's built on Argos Translate and can run entirely on your own machine.

Firefox's built-in translation — convenient if your main need is translating web pages. Its translation technology is designed to operate locally rather than sending the page to a remote translation service.

What I'd choose

If you tell me Windows, Mac, Linux, Android, or iPhone, and which languages you need, I can recommend the best specific option and give you the offline installation steps.