Product Updates *
Neural MT for all languages
All language pairs supported in Lilt are now powered by our Adaptive Neural MT technology.
The new Lexicon/concordance search is now a full-text search and contains meta information.
Automatic terminology lookup
When you open the Translate/Memory Search tab from the sidebar, your termbase entries are now also shown automatically for the given source text. Read more information about this in our knowledge base.
New languages added
Japanese, Korean, Traditional Chinese and Arabic are now available upon request. Lilt now offers interactive, adaptive neural MT for 6 languages!
Extended search for projects and memories
We have extended the project search so that additionally to search by name you can search also by language pair and due date (incl. today for projects due on the current date). Also memories are now searchable and you can search by name, language and memory ID.
The new Lilt
We have launched an entirely new Lilt! Our website and app got a new look and we launched some exciting new features:
- For businesses: The neural feedback loop - our adaptive neural machine translation system. Read more about this in our press release article.
- For everyone: A new editor interface for an improved translation experience. Read more about this in our Lilt Tips article.
New Project Report
The Business-edition project report now distinguishes between translation, research, and review time. For projects with teams, time tracking is more accurate for documents that are translated by multiple translators. Finally, the project report can be exported as a spreadsheet for offline analysis.
Speed / Stability Updates
Great translators work quickly. This means that Lilt needs to be fast, stable, and free of even the most minor annoyances. Speed and stability have long been priorities, but for the past four weeks we have focused on these issues. About 60 minor improvements and bug fixes have been released during that period.
New Project Details page
We have updated our project details page with a new design. Now documents and their details are arranged in a table and can be sorted per each column. The table can be exported as a detailed report.
Advanced Termbase Editor
Our new advanced termbase editor lets you manage terminology more effectively by keeping terms organized with meta information that you can customize. Import terminology with meta fields or add your own fields. Your terms will appear in both the Lexicon and the Editor suggestions and help you increase consistency and quality.
New and improved file filters
We updated and improved our IDML and SDLXLIFF file filters. Also we added new supported file types including Markdown (.md) and template and macros-enabled formats for MS Office (docm, xlsm, xltm, etc.).
Placeholders and non-translateable content
For text which must not be modified by the system in any way (segmented or translated) you can use the following placeholder format
This is especially useful for API users (but also available in the Web App) to handle names and placeholder formats of other systems.
Project-level find & replace
The Find & Replace functionality has been extended. Finally, you can search for terms and ensure terminological consistency across all documents in an entire project.
We have simplified the sign up page even more! Now you'll just enter your e-mail address and we automatically detect whether this is a Google e-mail address.
Our updated comments feature is better than ever! You can now edit and resolve comments, add clickable URLs (e.g. links to term definitions or images) and keep a thread history.
In-Editor QA Check
Our new editor includes extended proofreading features such as spellchecking, spelling correction and stylistic checks.
Business: Kanban Project Dashboard
Visualize the current status of your team's projects with the new Kanban dashboard.
Business: API Version 2
V2 of our API is now in beta for all Business users. The new API supports synchronization of large-scale TMs (i.e., 100M+ translation units) from TMSs to Lilt for MT training. Major features of the new API:
- Training of and translating with interactive, adaptive machine translation
- Large-scale translation memory
- The Lexicon (a large-scale termbase)
- Programmatic control of the Lilt CAT environment
- Translation memory (TM) synchronization
Annual pricing, Pro Plus
It turns out that people liked the $15 plan that we discontinued last November. It's back! We also offer annual pricing for both Pro plans.
14-day free trial of Pro and Business
Try the Pro and Business editions free for 14 days.
Business: In-app Word Block purchasing
Purchase word blocks through the app. The more you buy, the more you save.
Business: Share Projects and Memories with Managers
Project managers can now see the projects of other PMs. They can also see and use memories owned by other members of the team.
Faster scrolling in Editor, Optional auto-scrolling
Scrolling in the editor is now significantly faster. You can also disable auto-scrolling if you don't want the editor to advance auto-magically to the next unconfirmed segment.
Add your address, company name, VAT number, and other information to your account. These details will be added to your monthly invoices, the full history of which you can now see in the app.
Speed and stability are the themes for this month.
Chinese > English
Chinese > English is now supported (we had supported the other direction since the spring of 2016).
Faster web application
The Lilt web application has been fundamentally re-built. There's no need to upgrade: you'll get the latest software when you next login.
Bug fixes and design tweaks
To get on Santa's good list, we fixed 22 bugs and made 13 design tweaks. Improvements include a lowercasing hotkey, re-labeled buttons and icons, and a new tutorial, among many other small changes.
Automatic QA Checks
Run automatic QA checks in the editor QA mode:
The following automatic checks will be applied to each translated segment:
- Doubled words
- Doubled punctuation
- Doubled whitespace
- Leading/trailing whitespace
- Missing/inserted numbers (both source and target)
Live Terminological Consistency
Enforce terminological consistency in the machine suggestions. The Lexicon is now completely integrated with the adaptive MT system. Add a term either manually through the Lexicon window, or by uploading terminology to the custom Memory, and see the termbase entries highlighted in the MT output:
Use the terminology editor in the Memories view to add/edit/delete terminology entries.
The TM returns in-context (ICE) matches, also known as 101% matches.
Translation Memory Match Provenance
Translation memory matches now include the following information:
- Name of translator
- Email address
- Date confirmed
- Source document
Business: Revision History and Reports
Changes made by reviewers are now tracked. To export a revision report (in Microsoft Excel format), select Export > Revision Report in the project view:
For each source segment in the project, the revision report shows the original translation submitted by the translator, and the final translation accepted by the reviewer.
The behavior of the editor can now be configured at the account level. Disable exact match auto-propagation for a project? No problem. Go to Account > Options to disable the feature:
Add your translators and project managers to your Lilt team. The supply chain license means free accounts for everyone! Track performance over time.
Neural Morphology in the Lexicon
The lexicon now reduces queries using morphology rules induced with a neural morphology model. Search for an inflected verb, and see the citation form. Search for a plural noun, and see the singular form. No setup required.
100% Match Auto-propagation
Exact matches are automatically confirmed. A simple change, but one of the most requested productivity optimizations.
Pro: Referral Program
Pro users: refer a friend who upgrades to Pro and get a free month! Click the "refer" button at the top of the screen.
Polish and Norwegian are now publicly available. Russian will be available for testing this month. Lilt now offers interactive, adaptive MT for 11 languages!
Quote Builder and TM Leverage Report
Interested in going beyond per-word pricing? Our new Quote Builder lets you compute your effective hourly rate from your historical throughput and current TM leverage data.
Business: Comments and Project Reports
Generate project-wide reports for your translation team. Translators, reviewers, and project managers can also communicate through comments.
Improved Translation Quality
We've re-launched new baseline MT systems with the following translation quality (BLEU) improvements for each language pair:
- English-Spanish: Up to +2.5
- Spanish-English: Up to +2.0
- English-German: Up to +1.1
- German-English: Up to +2.6
Don't like our sentence segmentation? Split, merge, and split again until you're happy.
In-Editor Tag Manipulation
Don't like our automatic tag placement? Move tags until you're happy.
New File Filters
PDF (import only) and two new flavors of XLIFF: MQXLIFF (for memoq users), and MXLIFF (for Memsource users).
Baseline Translation Quality Updates
We've been improving our MT system. English>Chinese was the first to re-launch with up to 3.7 BLEU improvements on our internal test sets. We'll re-launch the other languages over the next few weeks.
Beta languages: Norwegian, Polish, Russian
Three new languages are now available for beta testing. Contact us if you want to help improve these language pairs.
(Business only) Assign a document for review. The reviewer can begin reviewing confirmed segments while the translator is still working.
Fuzzy Match Diff
The editor now shows the difference between the source side of a TM match and the current segment. Yeah, we should have added this feature long ago....
The editor now allows multiword auto-completion for sub-segments that come from a custom Memory.
Add terminology in TBX, Excel, CSV, or TSV format. Your terms will appear in both the Lexicon and the editor suggestions.
Translation Memory 2.0
We increased match recall on our standard test set from 86.2% to 95.5%, or a 9.3% improvement. No pre-processing is needed when you open a new document. Our TM scales to millions of entries instantly.
Large Document Support
The Editor now loads and paginates large documents much faster.
Custom Memory Export
Download a TMX file with all of your memory entries. Own your data.
Archive old projects to remove them from your dashboard and keep their data in your memory.
SDL Trados Studio Formats
We now support SDLXLIFF, SDLTM, and SDLPPX, i.e., Trados packages. Transfer data painlessly to and from SDL Trados Studio.
New languages: Chinese, Danish, Swedish
Did you know that the combined number of L1 and L2 speakers of Danish and Swedish is 1.6% of the number of Mandarin speakers? No matter. Lilt supports linguistic communities of all sizes!
Lightning Fast Editor
Translators often tell us that web-based tools aren't as responsive as traditional native tools. We know that lag is a productivity killer. For the last two months we've been rooting out sources of lag, the last of which was the typing component in the editor. Even very fast typists should find that the new editor can keep up.
After five months of beta testing, we officially launched Lilt! This revolutionary software allows you to optimize speed, cost, and quality at all points of the translation supply chain. There are three subscription tiers:
- Starter: free for evaluation and casual use
- Pro: for professional translators
- Business: for LSP and agency project managers
Project and document sharing (teams)
Upload a document and then assign it to team members. Group documents into shared projects for better project management.
Translation memory / machine translation sharing
Never before have teams been able to train an MT system collectively in real-time!
Create TM / MT memories for specific domains and clients.
Fast, scalable translation memory
See partial matches in the editor. Finally...and the wait was worth it. Lilt scales to millions of translation units instantly.
132% faster in Europe
We launched edge servers in Europe (Belgium) and Asia (Taiwan). On our standard benchmark, latency in Europe was reduced from 0.761 seconds per translation request to 0.328 seconds, or 132% faster.
New file filters
We added 18 new file filters including SDLXLIFF, Libre/OpenOffice, TTX, and CSV/TSV. We also updated our DOCX filter, which had caused a few sleepless nights.
Search and replace
Search for terms and ensure terminological consistency across segments.
We added English-Dutch and English-Portuguese.
A compact, tabular view for proofreading and quality assurance.