ncat-yan/ollama-translate-alfred — explained in plain English
Analysis updated 2026-05-18
Translate a word or sentence without leaving your keyboard, using a hotkey inside Alfred.
Look up an English word's Chinese meaning using the macOS dictionary or an offline dictionary.
Translate longer selected passages of text privately, without sending them to a cloud service.
| ncat-yan/ollama-translate-alfred | 0-bingwu-0/live-interpreter | 010zx00x1/faresnipe | |
|---|---|---|---|
| Stars | 2 | 2 | 2 |
| Language | Python | Python | Python |
| Setup difficulty | moderate | moderate | easy |
| Complexity | 2/5 | 2/5 | 2/5 |
| Audience | general | general | general |
Figures from each repo's GitHub metadata at analysis time.
Requires Alfred's paid Powerpack plus a local Ollama install and downloaded model.
ollama-translate-alfred is a workflow for Alfred, a macOS launcher app, that lets you look up English words and translate text entirely on your own computer instead of sending it to a cloud translation service. By default it translates other languages into Simplified Chinese and translates Chinese into English. When you type a single English word, it first checks the built in macOS dictionary, then an optional offline Chinese English dictionary called ECDICT, and only falls back to the local Ollama AI model if neither has an answer. For longer text, it sends the content to a local Ollama installation running a translation model called TranslateGemma, breaking long passages into chunks of about 1,200 tokens so they translate cleanly. You can trigger it a few ways: typing a keyword followed by a word or phrase, translating whatever is currently on your clipboard, translating text you have selected anywhere on macOS through a hotkey or through Alfred's Universal Actions, then copying, pasting, or viewing the full translation in a larger text view. Setup requires macOS, Alfred 5 with its paid Powerpack upgrade, and Ollama installed locally, plus downloading the translation model through Ollama itself. Settings like which model to use, the chunk size, and timeouts can all be adjusted inside Alfred's workflow configuration screen. The project includes its own build script, a test suite, and a written comparison of two different sizes of the TranslateGemma model. Because everything runs locally through Ollama, no text you translate is sent to an outside company, which matters if you are translating anything private. The project is released under the MIT license, with a separate notice file covering the license of the bundled ECDICT dictionary data.
An Alfred workflow for macOS that translates text and looks up English words using a locally running Ollama AI model, with no cloud service involved.
Mainly Python. The stack also includes Python, Ollama, Alfred.
Use freely for any purpose, including commercial use, as long as you keep the copyright notice.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly general.
This repo across BitVibe Labs
Verify against the repo before relying on details.