Deterministic verification
for every document an AI touches.
Built at Columbia. For work that has to be right.
Write with AI. Trust what it produces. ResearcherX builds a knowledge graph from your sources and flags contradictions in real time.
Each one makes the next more powerful. Watch them in action.
Drop PDFs, paste URLs, or write directly in the editor. MinerU parses layout-aware text, the chunker splits it semantically, and every segment is embedded for search.
Each document you ingest adds typed entities and natural-language relationships to a persistent Neo4j graph. The more you feed it, the denser and more useful it becomes.
Two seconds after you stop typing, every sentence is checked against your established facts. Red underlines for high-confidence conflicts, orange for likely ones. Click for the source, explanation, and full provenance chain.
Ask questions in natural language. The chat engine traverses your knowledge graph, retrieves relevant paths, and streams answers with source citations you can verify.
Literature reviews, multi-source papers, grant proposals. Catch when your draft contradicts your own cited evidence.
Cross-reference claims across depositions, statutes, and case law. Every assertion anchored to its source document.
API docs, research reports, policy papers. Keep long documents internally consistent as they evolve across revisions.
200k cloud tokens (~$1 of compute). After exhaustion, local inference via Ollama keeps the editor functional at zero cost.
Unlimited documents, cloud LLMs, full knowledge graph + chat, LaTeX export, multi-user collaboration.
Shared team knowledge graphs, SSO, on-prem deployment, bulk corpus ingestion, and API access to your team's graph.
Built at Columbia. For work that has to be right.