Skip to content
OctymPowered By Proofig AI
Security & trust

Your manuscripts stay private.

Confidentiality is core to how Octym is built, not an afterthought. The same commitment holds across the whole platform, however your review is configured. Here's how every manuscript stays private.

Never used to train models

Your manuscripts are never used to train AI models. Full stop.

Secure, private servers

Analysis runs on secure, private infrastructure. Confidentiality is core to how Octym is built.

Evidence-backed & auditable

Every finding is traced to its source, so every decision is reviewable and defensible.

Human-in-the-loop

Octym presents signals with evidence. A person always makes the final call, never an automated verdict.

FAQ

Security, answered.

No. User manuscripts are never used to train models. They are analyzed on secure, private servers and kept confidential.

Yes. Each finding is evidence-backed and traced to its source, so decisions are reviewable and defensible.

No one at Octym browses your manuscripts: review is automated, and staff cannot open documents in the ordinary course of work. Access to a specific document is exceptional, limited to a small number of authorized senior personnel, and happens when you or your organization ask Octym to investigate or assist with that document. Octym analyzes manuscripts and grant applications on secure, private servers and never uses them to train AI models. Confidentiality is core to how Octym is built.

Yes, Octym is built for confidential review, while a general-purpose AI chatbot carries real confidentiality risk for unpublished work. Octym analyzes manuscripts and grant applications on secure, private servers and never uses them to train AI models, and its AI providers are contracted on no-training terms too. Files are encrypted in transit and at rest, kept three months unless agreed otherwise, and deletable at any time. A chatbot cannot run image forensics; Octym also reviews the figures, statistics and citations, every finding traced to its source.

Octym keeps a manuscript and its review results for three months from submission unless agreed otherwise, and you can delete them at any time. Octym is a review platform, not a long-term storage system, so keep your own copy of any report you need. An organization that needs a different retention period can agree one with Octym.

Octym stores manuscripts encrypted, in transit and at rest, on Amazon Web Services in the United States or, for Australian-region accounts, in Australia. Stored documents are also backed up. Octym analyzes manuscripts and grant applications on secure, private servers and never uses them to train AI models, and its third-party AI providers are contracted on terms that prohibit training on what is sent to them. Confidentiality is core to how Octym is built.

The submitting party does. Whoever submits a document to Octym, a researcher, institution, publisher or funder, decides what review runs and is the data controller for the personal data in it. On an organization's account, the organization is the controller. Proofiger Ltd, the company that operates Octym, processes the document on the controller's behalf. Every sharing option is off by default, within controls set by your organization's administrator. Octym builds no persistent profile, risk score or misconduct record about any person beyond the review.

Not in identifiable form: Octym is not built for identifiable health information, so human-subject or patient-derived material must be de-identified before you upload it. A de-identified clinical manuscript gets the complete review. Octym reviews the entire paper across three categories of checks: internal manuscript integrity, external checks and requirements, and research quality. That includes checking that ethics approval and informed consent are disclosed. Like every upload, it stays confidential, and Octym never uses it to train AI models.

Review with peace of mind.

Contact us, or log in.