Frequently Asked Questions

We provide AI-powered language polishing and proofreading for academic manuscripts. Upload your paper, choose a service tier and AI provider, and receive a professionally edited document with tracked changes — typically in minutes, not days.

Upload an editable manuscript (.doc or .docx), select your service tier, AI provider, English variant (American or British), and style guide. Our system parses the document section by section and returns a Word .docx file with tracked changes. PDF is accepted only for Pre-Submission Review or supporting context.

Services that revise or transform a document require .doc or .docx; legacy .doc files are converted to .docx before processing. PDF is accepted as a primary manuscript only for Pre-Submission Review and as supporting context such as reviewer comments or proposal guidelines. Content deliverables are returned as .docx files. The maximum file size is 50 MB.

Each AI service supports up to 150,000 billable words in its primary file. Most journal articles, theses, dissertations, proposals, and reference lists fall within this limit. If the primary file is longer, split it into parts and submit them separately.

We bill the primary file for your chosen service; supporting context files are not added. In a structured Word document, billable content normally includes the title, abstract, headings, main text, captions, acknowledgments, appendices, and substantive notes. Author metadata, boilerplate, protected citation fields, and the reference list are normally excluded. References Formatting is the only service that revises and bills the reference list. Pre-Submission Review is the only service that accepts a PDF as its primary billable file, using the total extracted word count. PDFs accepted only as supporting context are not billable. Your exact billable file and count appear before payment.

Most manuscripts are completed in 5–15 minutes depending on length and service tier. Basic Edit is the fastest; Advanced Editing and Revise + Respond take longer because they perform deeper analysis. You can track real-time progress on the order status page.

You receive a Word document with tracked changes (insertions and deletions marked in revision mode). Open it in Microsoft Word or a compatible editor, review each change, and accept or reject as needed. Advanced Editing also includes inline comments explaining significant changes.

RevisePilot offers four frontier AI models: Claude Sonnet 5 and Claude Opus 5 (Anthropic), GPT-5.6 (OpenAI), and Gemini 3.8 Flash (Google). Claude tends toward conservative, precise edits; GPT-5.6 produces fluent rewrites; Gemini balances both. All four run with their highest-performance reasoning or thinking modes enabled. See our current model comparison: The Best AI Models for Academic Editing in 2026.

Yes. Our service performs language polishing — it corrects grammar, improves clarity, and refines academic tone without altering your research content, data, or conclusions. This is equivalent to what a human copyeditor does. Major publishers and journals permit language editing assistance. Your original ideas and findings remain entirely your own.

Our system is specifically designed to polish your existing text, not generate new content. The AI preserves your voice, structure, and arguments — it only improves language quality. Because the core content and ideas are yours, AI detection tools should not flag your work. We also instruct the AI never to write in an "AI voice" or insert meta-commentary.

Yes. Our system automatically detects inline citations (Zotero, EndNote, Mendeley, Word built-in), footnotes, endnotes, and cross-references. These are replaced with placeholder tokens before editing and restored to their original XML format afterward, ensuring your reference manager links remain intact.

RevisePilot uses four frontier models — Claude Sonnet 5, Claude Opus 5, GPT-5.6, and Gemini 3.8 Flash — from Anthropic, OpenAI, and Google. All four run with their highest-performance reasoning modes enabled. Claude tends to be conservative and precise, GPT-5.6 produces fluent rewrites, and Gemini balances the two. We periodically reassess our model lineup as new frontier models are released.

All four models are accessed directly through their providers' enterprise APIs (Anthropic, OpenAI, and Google Gemini), with all traffic originating from our United States backend. Your manuscript is encrypted in transit and at rest. Under each provider's enterprise API terms, your content is not used to train the underlying models.

All RevisePilot infrastructure — application servers, the database, and file storage — runs in United States regions on Google Cloud (Cloud Run, Cloud SQL, and Cloud Storage). LLM API calls go directly to Anthropic, OpenAI, and Google's enterprise endpoints. Your content is not used to train the underlying models.

Basic Edit focuses on grammar, spelling, punctuation, and academic style. Advanced Editing performs a comprehensive revision — it restructures sentences for clarity, improves logical flow within sections, enhances academic tone, and adds inline comments explaining significant changes. Advanced Editing costs 2 credits vs. 1 for Basic Edit.

Pre-Submission Review generates a simulated peer review of your manuscript — the kind of feedback you would receive from journal reviewers. It evaluates methodology, clarity, structure, and presentation, and provides constructive criticism to help you strengthen your paper before submission. No tracked changes are made; you receive a review document.

Upload your unrevised manuscript plus up to 5 reviewer comment files. The system maps each reviewer comment to the relevant section, revises your manuscript to address the feedback (with tracked changes), and generates a point-by-point response letter. You receive both the revised manuscript and the response letter.

We support academic manuscript translation from 8 languages into English: Chinese (Simplified), Japanese, German, French, Spanish, Russian, Portuguese, and Arabic. The translation preserves academic terminology and produces publication-ready English with tracked changes.

You can select from Chicago Manual of Style, AMA (American Medical Association), APA (American Psychological Association), Vancouver, or CSE (Council of Science Editors). If you choose "None," the AI applies general academic English conventions. The style guide affects formatting of citations, numbers, abbreviations, and other discipline-specific conventions.

Yes. You can use any service as a one-time purchase without subscribing — pricing is $2 per 250 words, with no per-service fees. However, subscription plans offer significantly better value — as low as about $0.35 per 250 words on the Lab annual plan, compared with $2.00 per 250 words when you pay per order.

Each subscription plan includes monthly credits. Every service costs 1 credit per 250 words (ceil), so pricing is based purely on your manuscript length. Subscription credits reset each billing cycle. You can also purchase top-off credits at $0.50 each — these never expire. When placing an order, subscription credits are used first, then top-off credits.

Yes. You can cancel anytime through the Stripe billing portal. Your subscription remains active until the end of the current billing period. Unused subscription credits expire at the end of each cycle, but any purchased top-off credits remain in your account permanently.

One-time AI-service orders, the $1 introductory offer, and top-off credits accept credit/debit cards (including Visa, Mastercard, American Express, and UnionPay), Alipay, and WeChat Pay through Stripe. Apple Pay, Google Pay, Link, or Klarna may also appear when available. Recurring subscriptions are card-only. All transactions are encrypted and PCI-compliant.

Yes. Your manuscripts are stored in cloud services with encryption at rest and in transit. Files are accessible only to your account. Our servers are protected by a cloud-based WAF, and all API communication uses HTTPS with strict security headers.

Absolutely. We do not share, sell, or use your manuscripts for any purpose other than providing the editing service you requested. Your unpublished research remains confidential. Your content is never used for AI model training. See our Privacy Policy for full details.

No. Language polishing and proofreading are widely accepted in academic publishing. Our service improves the language quality of your manuscript — it does not generate content, fabricate data, or write on your behalf. This aligns with COPE (Committee on Publication Ethics) guidelines and journal editorial policies that permit professional language editing assistance.