Raycast Easydict: A Multi-Provider Dictionary and Translator Inside Raycast
A Raycast Extension for looking up words or translating text in an easy way. Support 48+ languages, support Linguee and Youdao dictionary, support OpenAI, DeepL, Google, Bing, Apple, Baidu, Tencent, Volcano, Youdao and Caiyun translation.
At a glance
- What is it?
- Raycast Easydict bundles Linguee, Youdao, DeepL, Google, Bing, Apple, Baidu, Tencent, Volcano, Caiyun and OpenAI-compatible models behind one Raycast command. It is a strong fit if you already live in Raycast and want several translation sources at once, and a poor fit if you need offline lookup or a terminal workflow.
- Who is it for?
- Adopt Raycast Easydict if you already run Raycast daily and want dictionary detail plus several translation engines without leaving the launcher. Skip it if you need offline lookup, a CLI, or a Linux build, since the README only documents macOS and Windows.
- Can I use it commercially?
- Yes. MIT is a permissive licence: you can use, modify and sell software built on it, as long as you keep its copyright and licence notices.
- Is it still maintained?
- Yes. The repository last received commits 5 days ago.
- What is it written in?
- Mainly TypeScript, according to GitHub's language statistics.
Answers come from the project's GitHub data, last synced on September 15, 2026, and from our analysis. They are not legal advice.
DEEP OPEN-SOURCE ANALYSIS
The gap Raycast Easydict fills between a launcher and a browser tab
Looking up a word usually means leaving whatever you are doing: switch to a browser, open a dictionary site, paste the word, close the tab. Raycast Easydict collapses that into a Raycast command. The package.json declares four commands, and the main one, easydict, is titled Search Word. It takes an optional text argument, detects the input language, and queries your preferred target language. The extension is aimed at people who already use Raycast as their launcher on macOS or Windows and want dictionary-grade output, not just a single translation string. The README lists what a result can contain: basic translations, pronunciations, exam coverage, parts of speech, tenses, and web phrases. That is closer to a dictionary entry than to a machine translation box. The project is MIT licensed, originally by tisfeng and currently maintained by maxchang3, with ViGeng and litomore listed as contributors.
How providers, language detection and result ordering actually work
The architecture is provider-oriented. A separate Manage Providers command exists purely to order built-in providers and configure AI providers, which tells you the extension treats each translation source as a pluggable unit rather than hard-coding one engine. The README names the built-in set: macOS System Translation, DeepL, Google Translate, Bing Translator, Baidu Translate, Tencent Translate, Volcano Engine Translation, Youdao Translate, Caiyun Translate, plus Linguee and Youdao Dictionary for dictionary lookup. On top of that, LLMs from Raycast AI or custom OpenAI-compatible endpoints can be used for both translation and word lookup. The README states the extension automatically detects input languages and queries your preferred target language, and that results can be sorted manually. Two behaviours are worth knowing before you install. First, automatic querying of selected text is enabled by default, so the extension reacts to what you have highlighted. Second, system proxies are supported, which matters if some of these providers are unreachable from your network without one. The result is a fan-out query model: one input, several providers, merged output you order yourself.
Installing Raycast Easydict and running a first lookup
This is a Raycast extension, so Raycast has to be installed first. The README gives two paths. The store path is the install button on the Raycast Store page for isfeng/easydict. The manual path clones the repository and runs the dev script.
git clone https://github.com/tisfeng/Raycast-Easydict.git && cd Raycast-Easydict
npm install && npm run devAfter npm run dev, Raycast loads the extension in development mode and the Search Word command becomes available. For a normal lookup, invoke Search Word and type into the Query field, or pass text as an argument. The argument is optional and typed as text in package.json, so an empty invocation is valid and the extension will fall back to selected text. A second command worth wiring up early is OCR Translate, which is declared with mode no-view: it takes a screenshot, extracts the text, and queries it. The README also notes that Cmd + S plays word pronunciation manually, and that automatic audio playback of pronunciations is a supported feature. If you want to control which engines answer, open Manage Providers and reorder them there rather than editing files.
Where Raycast Easydict stops being the right tool
The extension is bound to Raycast. If your editor, terminal or CI pipeline needs translation, there is no CLI in the repository layout and no documented headless mode, so this is the wrong layer entirely. Platform coverage is another boundary. The 3.0.0 release notes say Windows support was added and that most features are adapted, which is not the same as full parity; the README does not enumerate what is missing on Windows, so treat any feature you depend on as unverified there. Linux is not mentioned. There is also a privacy dimension the README does not resolve. Selected-text querying is on by default, and the provider list includes third-party services and custom OpenAI-compatible endpoints, so text you highlight can leave your machine. The README does not document a per-provider opt-out for automatic selection, only the general provider ordering. Finally, the dictionary results depend on remote services such as Linguee and Youdao; if those change their responses, the rich entry format degrades in ways the extension cannot control.
Raycast Easydict compared with the standalone Easydict macOS app
The most direct alternative is the project's own standalone app, which the README links under a collapsed section titled Looking for the native macOS app. It is a separate repository, tisfeng/Easydict, described as a concise and easy-to-use translation dictionary macOS App. The difference is not the provider list so much as the integration surface. The standalone app can do things a Raycast extension cannot, and the README shows them: OCR screenshot translate and auto-select translate are presented as app screenshots and GIFs. The extension covers OCR through its own OCR Translate command, but the app owns the window, the global shortcuts and the selection behaviour. The trade-off runs the other way too. The extension inherits Raycast's argument passing, fallback commands and launcher speed, and it is the only one of the two with Windows support per the 3.0.0 release notes. If you want translation as one keystroke inside a launcher you already use, the extension is the lighter choice. If you want translation to be an application in its own right, the standalone app is the one the README points at.
Maintenance cadence, recent releases and what upgrading costs you
The repository is not archived, and the last push was on 2026-09-11, three days before this writing. Releases have been frequent: 3.0.0 on 2026-08-11 added Windows support, 3.1.0 on 2026-08-12 added Favorite Words, and 3.2.0 on 2026-09-10 added custom AI providers and Chinese stroke-order diagrams. That cadence has a cost. Favorite Words arrived in 3.1.0 and stroke-order diagrams in 3.2.0, so a user who installed before August 2026 has a materially different feature set, and the README does not document a migration path for saved data. The repository layout includes CHANGELOG.md, so release-level changes are traceable there rather than in the README. On licensing, the package.json declares MIT and the repository carries a LICENSE file. MIT is permissive, but it covers the extension code, not the providers it calls: each translation service has its own terms, and several of them have API keys or quotas that the extension does not manage for you. That is a terms question, not a code question, and it is outside what this review can settle.
Editorial conclusion
Adopt Raycast Easydict if you already run Raycast daily and want dictionary detail plus several translation engines without leaving the launcher. Skip it if you need offline lookup, a CLI, or a Linux build, since the README only documents macOS and Windows. Before relying on it, verify that the providers you care about still answer without an API key and that your selected text is being sent to the endpoints you expect.
Frequently asked questions
Does Raycast Easydict work on Windows?
Yes, according to the 3.0.0 release notes, which mention Windows support and say most features are adapted. The README does not list which features are missing on Windows, so verify the ones you depend on.
Do I need an API key to use Raycast Easydict?
Not for the default setup. The README says it works out of the box for looking up words and translating text, and custom OpenAI-compatible endpoints are optional. Individual providers such as DeepL or Baidu may still require their own credentials.
Which dictionaries and translation engines does Raycast Easydict support?
It supports Linguee and Youdao Dictionary for lookup, and macOS System Translation, DeepL, Google, Bing, Baidu, Tencent, Volcano, Youdao and Caiyun for translation. Raycast AI or custom OpenAI-compatible LLMs can also be used for translation and word lookup.
Can Raycast Easydict translate text from a screenshot?
Yes. The package.json declares an OCR Translate command with mode no-view, described as screenshot OCR that queries text from an image.
Does Raycast Easydict send my selected text to external services?
The README states that automatically querying selected text is enabled by default, and the provider list includes third-party translation services and custom OpenAI-compatible endpoints. The README does not document disabling automatic selection per provider.
How do I change which translation providers Raycast Easydict uses?
Use the Manage Providers command, which the package.json describes as ordering built-in providers and configuring AI providers. The README also notes that query results can be sorted manually.
Community notes