Privately Review Your Research
AI-Powered, 100% Local

Revyou runs entirely on your machine — no cloud, no API keys, no telemetry. Just you, your manuscript, and a rigorous AI review. Your manuscripts never leave your computer.

100% Local

Your LLM runs on your hardware — no cloud inference or uploads. Your manuscript is never sent anywhere to be reviewed.

No Data Leaves Your Machine

Manuscripts, reviews, PDFs, images — everything stays on your computer. Your research is never used to train external models.

You Own Your Reviews

Every review you generate belongs to you. Export to Markdown, DOCX, or PDF. No proprietary format lock-in.

Revyou dark theme reviewing a manuscript
Revyou light theme reviewing a manuscript
Drag to compare

Drag the slider to compare Dark and Light themes

How It Works

Three steps from manuscript to structured review.

1

Attach your manuscript

Drag-and-drop or browse for PDF, DOCX, ODT, PPTX, XLSX, images, or plain text files. Page images are rendered for vision-capable models.

2

Choose a review template

Choose from 3 free templates — full review, methods only, statistical audit — or unlock all 13 with Pro, including discipline-specific formats for biology, CS, physics, and more.

3

Get structured feedback

Receive a sectioned review with severity indicators, verified citations, and actionable recommendations. Export as Markdown, DOCX, or PDF.

Your Manuscripts Shouldn't Be Training Data

When you upload a research paper to a commercial AI service, it may be stored, analyzed, and used to train future models. Revyou is different.

With cloud AI services...

  • Manuscripts uploaded to unknown servers
  • Your data may train their next model
  • No control over retention or deletion
  • Monthly subscription fees per user

With Revyou...

  • Everything runs locally on your machine
  • Your research stays with you — always
  • You own every review output permanently
  • Free forever — Pro in early access

Features

Everything you need for rigorous manuscript review. Built entirely in Rust — fast, safe, and private by design.

Review Templates

Full review, methods audit, statistical analysis, and discipline-specific formats 13 templates · 3 free

Multi-Model Support

Run any GGUF model from HuggingFace — curated picks that fit your RAM, or your own repo/URL. Vision-capable models can analyze figures and diagrams

Conversation Management

Multiple conversations, global search, save and load sessions. Keep reviews organized by manuscript

Runs Locally

Runs models in-process via llama.cpp — no server, no Ollama, no add-ons. Your manuscripts, reviews, and data never leave your computer

CPU & GPU support

Runs on any modern PC — accelerated by your GPU when available, with automatic CPU fallback.

Pro features — included in your 7-day trial

Web & Academic Search

OpenAlex, Europe PMC, arXiv, PubMed, and Semantic Scholar tools let the model verify claims and find related work

Citation Verification

Automatically validates references against CrossRef and OpenAlex. Every citation gets a verified status

Export to DOCX & PDF

Export individual reviews or entire conversations to Word or PDF. Formatted and ready to share with co-authors

Vision Support

Renders document pages as images for vision-capable models. Choose text-only, hybrid, or full-visual mode

Questions

CPU & GPU support. 8 GB RAM recommended — smaller models run with less. Any GGUF model from HuggingFace.

Is it really private?

Yes. Models run on your machine and your manuscripts never leave it. The only network activity is downloading models, the optional search/citation tools (off by default), and a one-time license check.

Which models can I use?

Any GGUF from HuggingFace. Revyou recommends models that fit your RAM, or paste a repo/URL. Vision needs a model that ships a projector (mmproj).

How does Pro licensing work?

Pro is in early access and not on sale yet. It will be a one-time purchase (lifetime, a few devices). A free 7-day trial of the Pro features is included with the app.

Do I need Ollama or Python?

No. Revyou runs models in-process via llama.cpp — one app, no server, no add-ons, no command line.

Your Research Stays Private. Period.

Revyou runs models locally via llama.cpp — no cloud processing, no API keys, no telemetry. Every manuscript, every review, every document you attach stays on your computer. The only network activity is downloading models from HuggingFace, the optional search/citation tools, and a one-time license check.

Conversation data is stored in ~/.revyou/ with restrictive owner-only file permissions. On Windows: %LOCALAPPDATA%\revyou\. You can delete everything at any time from within the app.

Web and literature tools (OpenAlex, Europe PMC, arXiv, PubMed, Semantic Scholar) are optional and off by default. Citation verification uses only public DOIs.

Download

Free to download and use. Pro features are in early access, with a free 7-day trial in the app.

Linux
.tar.gz

x86_64 portable folder · GPU via Vulkan

Download
Windows
.zip

x86_64 portable folder · GPU via Vulkan · Windows 10+

Download
macOS
.dmg

Universal (Intel + Apple Silicon) · Metal GPU · macOS 12+ · unsigned (right-click → Open)

Download

Free vs Pro

FeatureFreePro
Built-in review templates3all 13
Custom prompts✓✓
Conversations3Unlimited
Export to Markdown✓✓
Web & academic search—✓
Citation verification—✓
Vision support—✓
Export to DOCX & PDF—✓
PriceFree foreverComing soon

Get notified when Pro launches

Pro is in early access. Leave your email and we’ll tell you the moment it’s available — no spam, unsubscribe anytime.

Version 0.3.6