Pre-publication review for manuscripts that matter

No — not unless you explicitly enable it. By default, Manuscripter runs entirely on your own server using offline analysis jobs. No manuscript text is transmitted to any external AI provider. Pro users can opt in to LLM-powered overlays (claims re-analysis, manuscript reasoning, hotspot assessment); these can be disabled individually or blocked globally with Zero Exposure mode.

Your manuscript is stored within your own infrastructure — the server you deploy Manuscripter on. Nothing is uploaded to a third-party cloud storage provider unless you configure one yourself.

Zero Exposure mode blocks all outbound connections from Manuscripter to external AI providers — both at the application layer (the server raises a 503 before any network call is made) and optionally at the network layer via an iptables denylist. When enabled, the 16 offline analysis jobs continue to function normally; the 4 LLM-powered Pro features are disabled.

The Audit Suite focuses on factual and legal risk: entities, claims, risk scoring, timeline extraction, contradiction detection, and living person flags. The Writing Suite focuses on craft and structure: grammar, repetition, pacing, editorial checks, date references, and writing assessment across chapters. Both suites run offline and are included on every tier.

Yes. The demo lets you enter the workspace immediately as a free, creator, or pro user without registration. The demo loads a sample manuscript — to analyze your own manuscript, you will need to create a free account.

The free tier includes the full Audit Suite (7 jobs) and Writing Suite (8 jobs) for up to 3 manuscripts. Exports are limited to Claims CSV. You can run up to 10 analysis jobs per day.

The Attorney Packet PDF is a Pro-tier export that compiles living person flags, risk scores, claims evidence, and timeline data into a structured packet for attorney review. It is not legal advice. It is designed for workflows where a legal reader needs to assess the manuscript data without accessing the workspace directly.

Analysis jobs are queued and processed asynchronously. The full offline suite on a 120,000-word manuscript typically completes within a few minutes on standard server hardware. LLM-powered jobs (Pro only) depend on API response times from the configured provider.

Manuscripter accepts DOCX and plain-text files. Support for additional formats is on the roadmap.

Yes. The 16 offline analysis jobs have no external network dependencies. Zero Exposure mode can be enforced at the infrastructure level, making Manuscripter suitable for deployment in environments with strict egress controls, including IRB-regulated research systems and legal team networks with NDA requirements.

Pro includes AI Claims Analysis (a second-pass LLM review of flagged claims), Manuscript Reasoning (theme and structural analysis using chapter embeddings), and Assessment Hotspot analysis. These jobs send selected manuscript segments to an external AI provider and require network access. They can be individually disabled or blocked globally by Zero Exposure mode.

There is no hard limit enforced by the software. Practical limits depend on your server configuration. Very large manuscripts (400,000+ words) may require extended processing time or additional server resources.

Multi-seat workflows are supported: multiple reviewers can comment and collaborate on a manuscript, and the review queue tracks contributor activity. For organisations requiring volume limits above the Pro tier, contact us about institutional pricing.