βœ“ Link copied
Private beta Β· invite only

From blank page to draft with 43 verified references, in 8 minutes.

GivePap reads your PDFs, your Excel files, your R scripts, and returns a structured scientific paper draft β€” Introduction, Methods, Results, Discussion, Conclusion + References with real DOIs. We don't write the paper for you: we give you back 3 weeks a month.

6
specialized AI agents
11
bibliographic styles
5
scientific APIs
EU
GDPR self-hosted

What you upload. What you get back.

You upload

  • πŸ“„
    Literature PDFs
    reference papers from your field, as many as you have
  • πŸ“Š
    Data tables (XLSX / CSV)
    raw and aggregated experimental data
  • πŸ“œ
    Analysis script (R / Python)
    lmer, emmeans, FDR, scipy, statsmodels…
  • πŸ–Ό
    Figures (PNG / PDF / EPS)
    boxplot, volcano, heatmap, any plot
A few minutes to a draft. The more starting material you give, the better it writes.

GivePap returns

A Markdown draft structured section by section, with citations verified via DOI lookup and honest [NEEDS DETAIL] tags where the experimental data is actually missing.

# Effects of Lactation Stage on Bovine Milk miRNA Expression
## Introduction
Recent advances in dairy science have highlighted the role of milk-derived microRNAs (miRNAs) as biomarkers of mammary gland physiology [Bobbo 2020]. Bovine milk contains complex miRNA profiles modulated by lactation stage, parity, and health status [Zecconi 2019]. Genome-wide approaches have identified...
## Materials and Methods
Mid-infrared spectroscopy was performed on 462 milk samples using [NEEDS DETAIL: spectrometer model]. Statistical analysis used linear mixed-effects models (lmerTest::lmer, FDR Benjamini-Hochberg correction, p<0.05) on log-transformed miRNA expression values...
## Results
F(9,452)=3.31, p<0.001 indicated a significant lactation stage effect...
## References
[Bobbo 2020] doi:10.1080/1828051X.2020.1784804 βœ“ verified
[Zecconi 2019] doi:10.3168/jds.2018-15665 βœ“ verified
[Dohoo 1991] doi:10.3168/jds.S0022-0302(91)78321-8 βœ“ verified
... + 40 more verified references
  • βœ“5 standard sections: Introduction, Methods, Results, Discussion, Conclusion.
  • βœ“References with DOIs verified via PubMed β†’ CrossRef β†’ OpenAlex.
  • βœ“Markdown, DOCX, PDF export β€” no lock-in, take the files and walk away.
  • βœ“Honest tags [NEEDS DETAIL] where the data is missing β€” it doesn't invent the gap.

6 specialized agents, not a single magic prompt.

ChatGPT with one prompt generates a page that looks like a paper. No coherence between sections. Hallucinated DOIs. Invented numbers. GivePap uses 6 specialized agents that cooperate β€” and check each other.

πŸ”
Researcher
reads the PDFs, generates a JSON research brief with 35–50 candidate citations
Sonnet
↓
πŸ“‹
Outliner
section-by-section structure, maps every claim to a citation
Polished
↓
✍️
Writer Γ— 5
one per section Β· Introduction Β· Methods Β· Results Β· Discussion Β· Conclusion
Mixed
↓
πŸ”’
Fact-Checker
reconciles every number in the draft against the actual uploaded Excel files
Polished
↓
πŸ“š
Citation Validator
chain: brief β†’ external β†’ PubMed β†’ CrossRef β†’ OpenAlex
DOI lookup
↓
🧐
Critic
simulates peer review, tags issues: blocker Β· major Β· minor
Opus

Claude reads your Excel. The Fact-Checker reads it back.

The standard objection against AI on papers is: "it invents the numbers." That's true for plain ChatGPT. It's not true here β€” because a second, independent agent reopens the file and re-verifies every figure in the draft against the source cell.

πŸ“Š source Β· dataset.xlsx Β· sheet "lactation_anova"
lactation_stageF_valuep_value
Early3.31< 0.001
Mid2.840.003
Late1.970.041
β†’
πŸ“ generated paper Β· Results Β§2

"ANOVA revealed a significant effect of lactation stage on miR-148a expression (F(9,452)=3.31, p<0.001), with early-lactation cows showing 2.3-fold higher levels than late-lactation cows."

No plausible-looking numbers. No hallucinated DOIs. No citations made to look real.

πŸ”’

Independent Fact-Checker

A second agent reopens the Excel files and re-verifies every number in the draft. Mismatch β†’ visible flag.

πŸ”—

Quadruple DOI lookup

Every [Author Year] passes through PubMed E-utilities, CrossRef and OpenAlex. If it's not there, it's flagged.

⚠️

Explicit [NEEDS DETAIL]

When the experimental data is actually missing (e.g. spectrometer model), the tag stays. No silent invention.

5 scientific sources. 11 bibliographic styles.

Every citation passes through a chain of external lookups. The final format is pluggable β€” the same pipeline produces JDS or Nature or IEEE without rewriting the paper.

πŸ” Scientific databases

PubMed E-utilities OpenAlex CrossRef arXiv Semantic Scholar

Fallback chain: if PubMed doesn't find a DOI, try CrossRef. If CrossRef also fails, OpenAlex. If none β†’ explicit flag on the citation, it doesn't get invented.

πŸ“š Bibliographic styles

JDS Vancouver APA Nature Cell ACS ACM IEEE PhysRev Elsevier Harvard Harvard generic

Style selectable at the run level. Change tier or change target journal, regenerate without rewriting β€” the formatter is separate from the writer.

GivePap isn't for you if...

Four cases where we tell you directly to look elsewhere. We'd rather lose a customer than sell them a tool they don't need.

Production-grade. Not a hackathon prototype.

Python 3.12 / FastAPI backend with 15,600 lines of code, 24 REST endpoints, 7 DB tables, Multi-agent pipeline in testing with partner researchers. Deploy via Docker Compose, data self-hosted in the EU.

🧠

Claude Opus 4.7

Primary model for the higher tiers. Sonnet 4.6 for the fast draft. Tier-aware routing.

πŸ‡ͺπŸ‡Ί

GDPR self-hosted EU

Data inside your perimeter. Anthropic zero-retention opt-out active on the pipeline.

πŸ—„

pgvector RAG-ready

PostgreSQL 16 + pgvector extension for semantic embedding on uploaded PDFs.

πŸ€–

Multi-agent pipeline

6 specialized agents, orchestration via Redis 7 + RQ worker.

πŸ“š

11 citation styles

JDS, Vancouver, APA, Nature, Cell, ACS, ACM, IEEE, PhysRev, Elsevier Harvard, Harvard.

πŸ“€

Export MD/DOCX/PDF

Three output formats. Markdown is the source, DOCX and PDF generated on the fly.

πŸ”“

No lock-in

Export the files at any time. The pipeline produces a draft, not a hosting service.

πŸ§ͺ

Testing in progress

We're running real tests with partner researchers. Public numbers will land after this phase.

🐳

Docker Compose

Single-file deploy. Next.js 15 + React 19 frontend, FastAPI backend, Postgres, Redis.

Want to try it on your real data?

Private beta open to 10 partner researchers β€” free for 3 months in exchange for structured feedback. Are you building a paper right now? Let's show you β€” we run the pipeline on your dataset, you see the draft, you decide if you need it.

πŸ“§ info@cra-srl.com