Frequently Asked Questions

AI Proofreading accepts supported PDF, DOCX, and plain-text files. The current free-access policy supports documents up to 50 pages. Image-only scanned PDFs may require OCR before their text can be analyzed.

A signed-in user with a verified email can currently submit up to five free proofreading jobs. After the allowance is used, or when a document exceeds 50 pages, the site directs the user to the Contact page.

Documents are queued so resource-intensive local engines do not overload the server. Your queue position is saved, and you may close the page and return later. Processing begins when a worker becomes available.

The local engine checks spelling, grammar, punctuation, spacing, agreement, selected tense patterns, sentence structure, clarity, academic style, terminology, and consistency. Some ambiguous findings remain suggestions rather than automatic changes.

Where available, you can download an editable proofread DOCX, a proofread PDF, and a proofreading certificate. Corrections show the replaced wording with strike-through and the corrected wording with highlighting.

No. Grammar, readability, citation, similarity, and AI-writing scores summarize automated findings. They are not guarantees of accuracy, originality, authorship, academic integrity, acceptance, or publication readiness.

The detector identifies passages whose model patterns exceed its configured threshold. It can produce false positives and false negatives. The result should be treated as an evidence signal and never as proof that a person did or did not use AI.

No. Similar wording may be correctly quoted, technical, unavoidable, common, or properly cited. Conversely, plagiarism can occur without a high text-overlap score. Review matched passages, sources, quotation, and attribution together.

Prism Nexus aims to preserve the original document structure and formatting, but complex layouts, unusual fonts, scanned pages, equations, floating objects, and conversions between PDF and DOCX can cause differences. Always inspect the downloaded file.

Files and extracted text are processed to provide the selected tools, store results, and support downloads or account history. Core language engines are self-hosted, while some research checks may use scholarly indexes. See the Privacy Policy for details.

Keep the document and error message, confirm that the required local services are healthy, and try once more. If the problem continues, use the Contact page and include the file type, approximate page count, stage, and visible error—without sending passwords or sensitive content.

No. It is an assistance platform. High-stakes academic, legal, medical, financial, publication, or employment documents may still require a qualified human reviewer and independent source verification.