Products
Two open-source tools that turn visually rich documents into knowledge graphs you can question.
Curate Labs builds knowledge automation. Two open-source tools do the work: Verdant gets knowledge out of documents, and GraphForge holds it in a graph you can question. Everything else on this site is a way to use them.
The tools
- Verdant — knowledge extraction from visually rich documents. Scans, tables, charts, and long PDFs become structured text, scored on your own pages. AGPL-3.0.
- GraphForge — an embedded, openCypher-compatible graph engine. Build a graph, ask a question, keep the evidence. No database server. Apache-2.0.
Verdant produces the text and structure. GraphForge holds the entities, relationships, and evidence that come out of it. Each is useful alone. Together they are a complete path from a document set to a checkable answer.
Three ways to use them
- Open Source — run both yourself, in your own environment, for free.
- Cloud — GraphForge Hub at graphforge.sh gives graph projects a stable address to publish, share, and clone.
- Services — we build the pipeline with you, on your documents, and you keep the code.
Supporting libraries
- XYG — the charting and graph visualization library we built to draw GraphForge results. Also useful on its own for very large datasets.
- Other libraries and research implementations are listed on the open source page.
Experiments
- startops.sh — a small experiment in shaping a new business around evidence. Not part of the knowledge automation stack.