About LitReview

An open repository of literature reviews written by scientists, assisted by AI.

What this is

LitReview is an open repository of literature reviews written by scientists with AI assistance, serving the basic sciences. Think of it as arXiv for community syntheses: every review has a stable identifier, an abstract page, and is available as rendered HTML and plain Markdown source.

The site is curated, not assembled by machines. Scientists write and submit the reviews; AI tools support them — drafting, synthesizing, and fact-checking — but humans own the text and take responsibility for it. A moderator reviews each submission before publication.

Why

Keeping up with primary literature is increasingly impossible for individual researchers. Reviews are among the most valuable scientific products — but today they are scattered across paywalled journals or trapped in private PDFs. LitReview gives reviews a permanent, open home with a citable identifier, so anyone can get an honest, human-authored overview of a field's state — written faster and more thoroughly with AI as a partner.

How reviews are made

  1. A scientist writes a literature review on a topic within an active area, using AI tools as drafting, synthesis, and fact-checking aids.
  2. They submit it through the public form with an abstract, keywords, and contact details.
  3. A moderator reviews the submission for suitability and approves it for publication.
  4. An identifier following the arXiv convention YYMM.NNNNN is assigned, and the review is published as HTML + Markdown.

Authors retain responsibility for their content. AI assistance is a tool, not a replacement for the author's judgment. AI use is declared openly on every review page — using agentic AI for drafting, synthesis, and fact-checking is the norm here, never a disqualifier.

Citable by design

Every review gets a stable arXiv-style identifier (YYMM.NNNNN), a permanent URL, citation metadata readable by Google Scholar, Zotero, and reference managers, and a suggested citation string on its page. All reviews are published under the CC BY 4.0 license, so they can be reused, shared, and cited without friction.

Areas

Write with the EcoSeek agent stack

LitReview is the archive for these reviews. If you want a factory to help you write them, pair it with EcoSeek — our open-source scientific agent environment. EcoSeek lets you deploy agentic AI assistants on your own hardware or lab cluster and use them for the kind of work that goes into a review: literature search (search_literature, query_papers), synthesis and critique loops via its Emily reasoning engine (DiDAL dialectical protocol, with automatic routing to Hermes for expert synthesis), and fact-chaining through 25 domain tools. The AIs it runs are your drafting and fact-checking aids — you still write, own, and publish the review here under your name.

Everything you need runs free and open source (GNU AGPLv3): local or cluster inference with EcoCoder, tool execution with EcoAgent, and the AgenticPlug gateway for multi-user lab and HPC setups with approval gates and audit trails. Same philosophy as LitReview: the scientist is the author, AI is the assistant — openly and without apology.

Open by design

All reviews are free to read and download. The site itself is open source at github.com/alrobles/litreview.org, and the agent stack that supports writing these reviews is open source too (EcoSeek).

Contact

A. L. Robles Fernández · Kansas Biological Survey, University of Kansas
[email protected]