Sci-database Documentation

Turn research papers into verified, structured databases — with every value traceable to its source.

Sci-database is an AI platform for researchers who need structured data out of the scientific literature: systematic reviews, meta-analyses, review papers, and machine-learning datasets. You search 200M+ papers (or upload your own PDFs), describe the data you need in plain English, and the Kateeb extraction engine (كاتب, "scribe") reads every paper and fills your database — then a second, independent AI model cross-checks the results so you can see exactly where the models agree and where they don't.

The workflow at a glance

 ┌─────────────┐   ┌──────────────┐   ┌─────────────┐   ┌──────────────┐   ┌─────────────┐
 │  1. SEARCH   │ → │  2. VERIFY   │ → │  3. SCHEMA  │ → │  4. EXTRACT  │ → │ 5. VALIDATE │
 │  200M+ papers│   │  PDFs fetched │   │  AI suggests │   │  Kateeb reads │   │  AI Judge    │
 │  or upload   │   │  & checked    │   │  you edit    │   │  every paper  │   │  cross-checks│
 └─────────────┘   └──────────────┘   └─────────────┘   └──────────────┘   └─────────────┘
                                                                                   │
                                              ┌────────────────────────────────────┘
                                              ▼
                                   ┌─────────────────────┐
                                   │ 6. EXPORT & SHARE    │
                                   │ Excel · CSV · JSON   │
                                   │ or publish/sell it   │
                                   └─────────────────────┘

Papers in. One verified spreadsheet out. Zero synthetic data — if a value is not in the paper, the cell says N/A.

Documentation contents

1. Introduction — what Sci-database is, who it's for, and the design principles that make its output trustworthy.
2. Getting started — sign in, get your free credits, complete your profile.
3. Searching for papers — OpenAlex, CORE, and Semantic Scholar; filters, API keys, and the queue.
4. Files & uploads — automatic PDF verification, uploading your own PDFs, fixing papers that can't be fetched.
5. Building your schema — describe your database in plain English, let the AI draft the fields, then edit and save.
6. Running an extraction — processing jobs, statuses, stop & resume, result emails.
7. Validation & the AI Judge — cross-model checking, investigating disagreements, and converging above 90% agreement.
8. Your database & exports — viewing results and exporting to CSV / JSON / Excel.
9. Sharing & the marketplace — public schemas, free sharing, and listing a "golden" database for sale.
10. Credits & pricing — every credit cost in one table.
11. FAQ — the questions researchers actually ask.
12. Troubleshooting — every error message and how to fix it.

Developed by Dr. El Tayeb Bentria at Hamad Bin Khalifa University (HBKU), Doha, Qatar · Patent QF Ref: 2024-066.