DETAILED FAQ
Scores and evaluation questions, explained
Decision metrics, product fit, language, integration, security, and human-review guidance for buyers and pilot teams.
What does the Document Analyzer recommendation mean?
Follow the Recommended Translation Route: prepare the source first; AI-only for low-risk internal-reference use; AI translation plus bilingual human proofreading; human translation plus independent proofreading; or specialist human translation and sign-off. The route combines intended use with Translation Workflow Readiness, Production Risk, Translation Difficulty, content type, terminology, protected items, file format/layout, locale-sensitive data, and extraction reliability. Workflow readiness is a positive routability score—not translation quality or permission to skip the stated safeguards. Risk and difficulty are separately calculated higher-is-worse or higher-is-harder metrics and do not need to total 100 with readiness. Analysis Confidence states how strongly the evidence supports the assessment, not whether the document is safe.
When does Lingivo recommend AI-only or human translation?
AI-only is limited to low-risk, low-difficulty internal-reference use with strong extraction and analysis confidence; it must not be published or delivered. AI plus bilingual human proofreading is the normal route when AI can prepare a draft but a qualified reviewer must approve it. Human translation plus independent proofreading is used for high-risk or high-difficulty work. Specialist human translation is used for critical, legal, medical, regulated, or contractual use and adds required subject-matter sign-off. Weak or insufficient extraction stops translation until the source is repaired and reanalyzed.
What does Translation Draft Confidence mean?
Draft Confidence measures whether the generated draft is technically ready to enter bilingual LQA. It combines segment coverage 35%, structural preservation 25%, approved terminology compliance 15%, translation completion 15%, and provider/segment confidence 10%. Scores of 90–100 are ready for LQA, 82–89 need draft review first, and scores below 82 need repair before LQA. Number-format changes that may be valid for the target locale are review warnings rather than automatic blockers. This is not a linguistic-quality or delivery score.
How should I interpret the LQA Quality Score and delivery risk?
The LQA Quality Score is normalized across the categories selected for the run. Only located rule evidence and AI findings with exact validated source/target segment evidence and confidence of 80% or higher can lower a category score; lower-confidence AI observations remain non-scoring review notes. No validated evidence means no penalty. Below 65 recommends human retranslation, 65–89 requires human review, and 90+ requires the listed corrections and approval. Delivery risk is not simply 100 minus the score: a critical Grammar, Terminology, Consistency, or Contextual finding can block use even when the weighted total is high.
Why is there no overall Glossary score?
A glossary is a governed terminology artifact, not a pass/fail output. Lingivo shows confidence and termhood for individual candidates and keeps every proposed translation unapproved until a linguist, reviewer, or authorized workspace user accepts it.
Why is there no overall Style Guide score?
A generated style guide combines confirmed client rules with inferred guidance. Lingivo keeps those sources separate so reviewers can approve the enforceable rules that Translation and LQA may reuse; a single percentage would hide that governance distinction.
Does Lingivo replace human translators or reviewers?
No. It accelerates project preparation, controlled translation drafts, and issue detection while keeping professional reviewers responsible for judgment and final approval.
How does Lingivo compare with CAT tools?
Lingivo complements rather than replaces CAT tools. It adds intake analysis, terminology and style preparation, governed AI translation, evidence-backed LQA, and reusable exports around the CAT workflow.
Does Lingivo integrate with CAT tools, APIs, or enterprise systems?
Lingivo has an application API and structured exports. Public API documentation and direct CAT, automation, support-platform, and marketplace connectors are not yet published; confirm required integrations during the pilot.
Which languages and document types are supported?
Lingivo provides more than 120 searchable language and locale options. The app accepts DOCX, PDF, PPT/PPTX, XLIFF, SDLXLIFF, MQXLIFF, CSV, JSON, XML, PO, XLS/XLSX, TBX, and SRT files up to the configured upload limit.
Will the translated download keep the source format?
Lingivo returns the same format for supported editable files only after the export is verified. PDF layout preservation needs additional validation and may require an alternate export.
How does Lingivo handle low-confidence language pairs?
Confidence messaging considers language-pair coverage, extraction quality, references, script requirements, and content risk. Medium or low confidence explicitly requires human or specialist review.
How does Lingivo handle uploaded files securely?
Uploads are validated, fingerprinted, held outside the public webroot, processed under organization access controls, and removed according to retention policy. Lingivo does not use them to train AI models.
Where can I review trial limits and credit pricing?
The Pricing page contains role-specific plans, the full credit calculation table, trial terms, add-on packs, and pricing FAQs.
