Atlas
Atlas: every lab has a taste.
Core trick
Builds a semantic-search index over a lab's deduplicated papers and exposes it via MCP so Claude checks idea novelty against owner-controlled memory.
Description
Atlas: every lab has a taste.
AI threatens to homogenise scientific research: every lab asks the same models to run its literature search, summarise papers, and come up with ideas - and every lab gets the same consensus answer. Atlas is the bet against that. It doesn't make Claude smarter; it gives Claude your lab's perspective.
Atlas builds a shared, searchable memory from a lab's literature: deduplicated papers, semantic search, topic maps, and grounded digests of what's new. As researchers read, discuss, and curate papers, Atlas learns the lab's evolving scientific taste.
Through MCP, Claude can securely read that memory (owner-controlled and fully audited) and answer a question a generic literature search can't: is this idea novel for our lab? It drafts related-work sections in which every citation is verifiable, and even generates figures in the visual style the lab has implicitly developed over years.
Instead of every lab converging on the same AI, Atlas gives every lab an AI shaped by its own scientific taste.
https://atlas-papers.vercel.app/papers
README excerpt
<div align="center"> # Atlas **Every lab has a taste. Atlas gives yours to Claude.** The papers you save, the figures you admire, the ones you argue about in group meeting — Atlas learns your lab's judgment from how you already work, then hands it to Claude. [**Live app**](https://labatlas.app) · [API](https://paper-radar-api.fly.dev/docs) · [MIT licensed](LICENSE) </div>  --- ## The problem A lab's taste — what's worth reading, what's overhyped, what a good figure looks like — lives in one senior postdoc's head. When they finish, it walks out the door. What's left behind is a Dropbox folder and a dead Teams channel. Atlas keeps it. It turns the papers a lab already shares into a corpus that is searchable, mapped by meaning, and — crucially — **legible to an LLM**. Claude can then answer the question a literature search can't: not "what exists?" but "what's new *to us*?" ## What it does - **Shared corpus.** Papers posted by URL are deduped, resolved against arXiv / Crossref / PubMed / Europe PMC, enriched with summaries and tags, and made searchable by keyword or by meaning (Voyage embeddings + pgvector). - **Learns from behaviour.** No profile to fill in. Saves, reactions, and comments are weighted differently — a save counts more than a 🤔 — and decay over time, so recommendations track where the lab is going, not where it's been. - **Topic maps.** A map is a topic the lab watches, kept current: papers cl