The question that keeps coming up in language forums and group chats is some version of this: my go-to site for real book examples just vanished, and now every translation feels either too stiff or too guesswork. The site in question, tr-ex.me, delivered something many tools still struggle with: phrase translations pulled straight from published books and their official versions in other languages, complete with full sentence context.
Users valued it because a single search for an idiom like “take with a pinch of salt” returned not only several natural equivalents but also the surrounding paragraphs from novels where those equivalents actually appeared. That level of sourcing helped learners and professional translators alike see register, tone, and collocations in action rather than isolated dictionary entries.
The database drew from literature, film subtitles, and transcripts, which gave it an edge over purely statistical or corporate-text collections. Language enthusiasts working between English and Czech, for instance, repeatedly noted how the examples captured everyday or literary “sillyness” that drier corpora missed. The site operated for years as a quiet favorite among serious learners who wanted proof from printed sources instead of machine suggestions.
Photo by Jason Leung on Unsplash
Its disappearance happened without a formal announcement. Visitors began receiving a 403 Forbidden error, and the last weeks before the outage featured heavier mobile advertising, which many interpreted as a final attempt to cover server costs. Maintaining a large, high-quality parallel corpus of book translations carries ongoing expenses for storage, scraping compliance, and legal review. Copyright questions around aggregated literary material likely added pressure. At the same time, the rapid adoption of large language models reduced traffic for many niche reference sites, making continued operation harder to justify.
Users expressed genuine disappointment rather than anger. Several described daily reliance on the tool for professional work or personal study and noted the sudden gap it left. One Czech translator mentioned using it as a quick reference even in paid projects. The absence of any update from the operators left the community piecing together the likely causes from patterns seen with similar projects.
Reverso Context stands out as the closest current option for many. It pulls example sentences from a wide range of real texts, including literature, news, and subtitles, and presents them side by side with the original. The interface lets users filter by domain or register, which helps when the goal is natural phrasing rather than literal accuracy. For pairs that include less common languages, coverage varies, yet the volume of user-submitted and verified examples continues to grow.
Linguee, now closely tied to DeepL’s ecosystem, offers another strong contextual layer. Its strength lies in professional and technical documents, so the examples tend toward formal or specialized usage. Learners seeking casual or literary idioms sometimes find the results narrower than tr-ex.me once provided, but the reliability of the sources remains high and the offline mode proves useful on the move.
WordReference continues to serve as a dependable dictionary backbone with active forums where native speakers discuss tricky phrases. It lacks the massive pre-populated book corpus, yet the community threads often supply the exact contextual nuance a user needs. For targeted language pairs such as English-Czech, specialized corpora like the one at korpus.cz deliver searchable context from Czech national sources and paired translations through its Treq tool.
Ludwig Guru and Sentence Stack focus more narrowly on English writing improvement, pulling examples from published work to show correct usage. They complement rather than replace broader multilingual tools. Open-source or community-driven projects occasionally surface as people attempt to rebuild similar databases, though none have yet matched the scale or curation quality of the original site.
Choosing among these options depends on the specific need. For quick phrase checks with literary flavor, start with Reverso Context and cross-reference Linguee when the material leans professional. When deeper verification matters, combine a dictionary forum with a national corpus site for the target language. Many learners now keep two or three bookmarks open and rotate based on the pair and the register required.
The team that kept tr-ex.me running quietly performed the kind of steady curation that rarely receives public credit. They assembled and maintained a resource that rewarded careful users with authentic examples no automated system fully replicated at the time. Their work demonstrated that good process, applied consistently to data collection and presentation, creates lasting value for the people who rely on it.
One practical step a regular user can take this month is to test Reverso Context on three idioms or phrases that previously sent them to tr-ex.me and note which examples feel closest to the old results. That short trial usually clarifies whether the new workflow fits or whether a second tool needs adding to the rotation.
