Platform · Quality Assurance
Catch Every Error
Before Your Client Does
More than two dozen configurable rule categories run right inside the editor, covering tags, numbers, terminology, length, spacing, and consistency. Issues land in a review panel beside the segment, and a mandatory gate can block job completion until the critical ones are resolved. Quality stops being a final scramble and becomes part of the work.
The Check Library
One Toolkit, Every Common Defect
Wordbee ships with more than two dozen rule categories out of the box. Most are on by default; a few stay off until you need them. Each one is configurable, can be set to error, warning, or information, and runs the same way whether a linguist triggers it or a workflow step requires it.
Tags and Markup
Finds missing, extra, or misplaced formatting tags and verifies duplicated tags carry across. Critical for HTML, XML, and InDesign, where a broken tag corrupts the output file.
Numbers and Dates
Confirms every number and date in the source survives into the translation and is localized to the target region. Catches a dropped decimal or a date in the wrong format before it ships.
Terminology
Checks approved terms from your termbases against a configurable similarity threshold, and screens for forbidden terms from a list you define per profile.
Consistency
Flags identical source segments translated differently, different sources translated the same way, untranslated content, and segments left identical to the source.
Spacing and Punctuation
Leading and trailing spaces, double spaces, spaces before punctuation with language-specific exceptions, doubled punctuation, and segment-boundary spacing for CJK pairs.
Length and Patterns
Enforces character or percentage length limits for UI strings, runs custom regex patterns over source and target, and includes spell check, repeated words, and sentiment.
QA Profiles
Set the Rules Once, Reuse Them Everywhere
A QA profile is a named set of rules with a severity assigned to each. Start from the default profile, clone it, and tune it for a client, a domain, or a quality tier. The same profile drives a manual check in the editor and a mandatory gate on a workflow step, so a project is held to exactly the standard you defined for it.
- Default profile enables most rules out of the box; clone and customize from there
- Three severity levels: error must be fixed, warning is reviewed, information is a notice
- Scope a profile to specific source and target languages
- Configure an automatic action per rule: change segment status or set a bookmark
- Skip green-status, locked, or already-validated segments to cut the noise
- Keep strict profiles for regulated clients and lighter ones for high-volume work
QA as a Workflow Gate
The Same Loop on Every Job
QA is not a one-time pass at the end. It is a loop a step can require: configure the profile, run the check, review and resolve, then prove it. Make it mandatory and a supplier cannot complete the job while errors remain.
- Re-run after each edit; resolved issues drop off and the count only falls
Step 1 · Configure
Configure the Profile
Pick or build the QA profile that fits this project, with the rules and severities that matter for the client and content type.
At this step
- Clone the default profile and adjust per client or domain
- Set each rule to error, warning, or information
- Assign the profile to the relevant workflow step
Linguistic Quality Assessment
For the Quality a Rule Cannot Score
Automated checks catch mechanical defects. They cannot tell you whether a translation reads naturally or hits the right register. LQA adds a structured human review on top, scoring work against a defined model so quality becomes a number you can compare across languages and suppliers. The same exportable record proves the review happened: attach it to a delivery or hand it to an auditor.
- Annotate by error category (accuracy, fluency, terminology, style, and locale), each weighted by severity
- Produce a single score against a pass threshold you set per client
- Mirror established frameworks such as MQM or DQF so scores stay comparable across suppliers and language pairs
- Roll scores up per document, language pair, or supplier to see where quality dips and which vendors hold up
- Sample a representative share of large jobs instead of reviewing every segment
Common Questions
What Teams Ask
See a QA Profile Built for Your Content
Talk with our team about the rules, severities, and delivery gate that fit your clients and the report your buyers will see, or start free and build a profile yourself.