A luminous library hall where research paths converge

Research intelligence

LitSynth

A systematic-review workspace that turns scattered papers into evidence you can cite with confidence.

PRISMA-nativeLocal-firstProvenance citations

The idea

Research, synthesis, understanding.

Modern research is abundant and fragmented. The papers pile up; the connections between them stay invisible. LitSynth treats knowledge as a network — and builds the tooling to screen, connect and cite it without losing the thread.

Built for PhD students, systematic reviewers and anyone writing a related-work section that has to hold up.

A manuscript glowing with connected annotations

How it flows

From documents to understanding.

  1. 01

    Data

    Bring your library in — RIS/BibTeX, PDFs and notes from Scopus, PubMed, IEEE and arXiv. Everything is indexed into one local-first workspace that stays on your machine.

  2. 02

    Search

    Move through PRISMA phases — identification, deduplication, screening, eligibility — with semantic search that finds the papers keyword search misses.

  3. 03

    Ask

    Chat with your library through evidence-grounded RAG. Every answer carries provenance you can trace back to the exact paper and page.

  4. 04

    Analyse

    Extract evidence into review matrices, compare studies side by side and let automated extraction surface what matters across hundreds of records.

  5. 05

    Write

    Draft with citation-safe assistance and export PRISMA reports, review matrices, DOCX, PDF, CSV and JSON — ready for submission.

The LitSynth pipeline glowing through a research workspace

PRISMA-native reviews

The full systematic-review pipeline — research question to inclusion — with deduplication and evidence extraction built in.

Local-first & private

Your library and index live on your machine. Run fully offline, or mirror to the cloud when you choose.

Provenance citations

Deterministic citations backed by PDF metadata — no hallucinated sources, ever.

AI-assisted drafting

RAG-generated text grounded in your papers — the assistant writes with your sources, not instead of them, and shows its work.

Evidence-grounded chat

Ask questions of your indexed papers and get answers with source attribution you can defend.

Bring your own model

Works with Ollama locally, or OpenAI, Claude, Gemini and Qwen — your data, your choice.

Multi-format export

PRISMA reports, review matrices, DOCX, PDF, CSV and JSON for submission.

Cross-platform

Desktop, web and tablet through one Flutter workspace.

Free for researchers

Auditable and built for researchers by a researcher — no paywall between you and your own evidence.