Knowledge automation, open source
Turn documents into knowledge you can question.
Verdant reads the scans, tables, and long agreements other tools skip. GraphForge turns what they contain into a graph your team and your agents can query, with the evidence attached. Run both yourself, use the hosted Hub, or have us build the pipeline with you.
How it works
One stack. Three ways to run it.
Verdant and GraphForge are open source and work on their own. Use them free, publish and share graphs through the hosted Hub, or have us put them to work on your documents. The code is the same in every case, and your documents stay in your environment.
Run GraphForge and Verdant yourself. Apache-2.0 and AGPL-3.0, with nothing held back.
GraphForge Hub gives your graph projects a stable address to publish, share, and clone.
We build the pipeline with you, on your documents, in your environment. You keep the code.
Extract with Verdant
Turn visually rich documents into structured Markdown and HTML, with charts rebuilt as diagrams and every run traced and scored.
Connect in GraphForge
Load entities and relationships into an embedded, openCypher-compatible graph with Arrow results and Parquet persistence.
Ask, with evidence attached
Rank, cluster, trace paths, and search through an agent or your own code. Record claims, their evidence, and your confidence.
The tools
Two open-source tools that meet in the middle.
Verdant gets knowledge out of documents. GraphForge holds it in a graph you can question. Each is useful on its own. Together they are a complete path from a pile of PDFs to a checkable answer. Below the tools are the teams we build for.
Extract
Knowledge extraction from visually rich documents.
Verdant digests scans, tables, charts, and long PDFs into structured text that people and agents can use, and measures the result on your own pages.
Connect
A graph engine with no server to run.
GraphForge is an embedded, openCypher-compatible graph engine for research and investigative work. Build a graph, ask a question, keep what you learn.
Data and AI teams
Pilots that stall on scanned PDFs and tables, fixed in your own environment.
Investigations and due diligence
Case graphs that keep the evidence behind every answer.
Agent and retrieval builders
An embedded, Apache-2.0 graph behind your agent or retrieval system.
Research and heritage
Free tools and published graphs for research groups and collections.
Start small