Documents and notes decision
ReadCube Papers
A competent developer can reproduce a useful subset (import/search/QA over a personal library) for about a week of work plus modest monthly API/host costs, but you cannot replicate ReadCube's proprietary corpus, publisher integrations, enterprise features, or brand-scale support without vendor access.
Visit website↗Open-source builds that already do this
Every project below is open source and already does this job today. Fork one, self-host it, or take the parts you need - the build prompt further down assumes an empty file, and this is the shortcut past that. Licences differ; check the one on each card before you ship. All ReadCube Papers alternatives, with the arithmetic →
What a replacement has to do
- Import PDFs and metadata, index full text + metadata, render/annotate PDFs in-browser, provide search and saved-alerts, and a question-answering chat over a user's library using embeddings
What it still won’t have
- Access to ReadCube/Dimensions proprietary indexed corpus and document-delivery integrations
- Publisher-specific AI permissions and the AI Access Indicator
- Enterprise features: SSO, admin dashboard, dedicated support, and custom branding
- Scale, reliability, and polished cross-device sync achieved by the vendor
- Out-of-the-box PRISMA/systematic-review workflows and dual-review automation
What remains hard
- Proprietary data
Manage your research, or ask a question, and get answers drawn from 165M+ publications and 182M+ patents.
- Proprietary data
Access 123M+ indexed papers from the Dimensions research database.
- Brand trust
Hundreds of companies, from startups to global leaders, trust ReadCube’s award-winning platform to transform how they work with scholarly literature.
First-year cost
Keep paying
Paying is—cheaper in year one.
On cash alone, building overtakes the subscription at 23 seats.
Money you would actually spend
Time you would spend
—
What you would spend
What we assumed
The verdict above measures whether you could build it. This one is only about money.
Runnable build prompt
Build a self-hosted minimal literature manager using: React (web UI), PDF.js for viewing/annotations, Node.js/Express backend, Postgres for metadata, S3-compatible storage for PDFs, and a vector DB (e.g., RedisVector or Faiss) plus Elasticsearch/Meilisearch for keyword search. Implement: PDF import with metadata extraction and EndNote/Zotero import; per-file full-text extraction and embedding generation (OpenAI or compatible embeddings API); keyword + semantic search; in-browser PDF viewing and highlight/annotation persisted to Postgres; an AI Q&A endpoint that retrieves top-k docs, constructs a context window, and calls an LLM API to return cited answers; user accounts and 1-tenant shared libraries. Out of scope: enterprise SSO, publisher paywalled document delivery integrations, and proprietary Dimensions corpus. Include error handling, rate limiting for API calls, and automated tests for ingestion, search, and the QA API.
How we checked
How the score was reached
- Partly verdict base52
- An open-source build was found+5
- 4 cited sources+3
- Price verified on pricing page+3
- Hard moats found in the evidence-3
- Evidence score60
The base comes from the verdict. Everything under it is a check that either happened or did not, and each one is a fact frozen in this record rather than a judgement made at render time - so the same evidence always produces the same number.
How scoring works →Cited sources · 4
Every page the run actually retrieved.
- official productReadCube product overview
- official pricingReadCube pricing
- official productReadCube for R&D Teams
- open sourceJabRef/jabref
Integrity checks
What held up, and what did not.
