Trust

Methodology and trust

AutoSearch is designed as a research assistant, not an authority. Every answer should remain traceable to sources you can inspect.

12 scientific sources OpenAlex, Crossref, PubMed, arXiv, Semantic Scholar, ClinicalTrials, EUR-Lex, ORCID, DBLP, DOAJ, Espacenet, Unpaywall

1.5M knowledge-base chunks Internal indexed RAG memory for retrieval before synthesis.

24 min deep-review paper Typical deep-review preset for manuscript-grade synthesis.

5 languages English, Italian, German, French, and Spanish outputs.

100% DOI verification Every cited DOI is checked through Crossref before export.

5 LLM providers MiniMax, DeepSeek, DashScope, Claude, and GPT planned.

Scientific research standards

PRISMA 2020 framing for transparent search, screening, and reporting discipline.

COPE-aware publication ethics: provenance, conflicts, and correction-friendly wording.

DOI and ORCID signals help connect papers, authors, and persistent scholarly records.

OpenAlex, Crossref, and OpenCitations are used as source graphs where coverage is available.

4-stage workflow: Ingest -> Index -> Retrieve + Synthesize -> Cite

Ingest files, URLs, and questions with metadata and safety checks.

Index extracted text and bibliographic signals for retrieval.

Retrieve relevant passages, synthesize cautiously, and separate evidence from inference.

Cite sources with IDs and export-ready references.

Bias clause

AI can miss sources, over-weight visible literature, or summarize incorrectly. Use AutoSearch for acceleration, then verify sources, methods, and citations before academic or operational use.

Sample dossier

Open sample dossier

Built for responsible evidence synthesis

We re-verify every DOI cited in real AutoSearch deliverables against Crossref / doi.org with an independent, re-runnable harness (scripts/benchmark_citations.py). Results are published as measured — including failures.

Trust & transparency →