Dependencies
What would a change affect?
Follow applications to servers, or suppliers to the customers that depend on them.
VioGraph · Self-hosted graph intelligence
Find answers in connected data.
Explore your data as a graph. Trace dependencies, review access and inspect the records behind each finding.
The platform and its graph data run in your environment.
An analyst’s journey
Your environmentSign in
Analyst
Connect data
Map relationships
Choose engine
Build graph
Explore graph
Inspect records
Evidence
SSO · Admin console
Ask AI · Optional01 — What it helps you find
What could you find?
Access reviews
Who can access this information?
Follow people and groups to the documents they can access.
Connected records
Which records are connected?
Find shared identifiers and relationships across records in separate tables.
02 — How it works
From your data to a working graph.
Connect your data
Upload files or connect a database, cloud storage or content source.
Map the relationships
Define nodes and connections, or import a reusable VioGraph schema.
Choose an engine and build
Choose an administrator-enabled engine and build from a saved copy of the data.
Explore and follow changes
Explore in the UI or query with code. Save findings and watch changes after refresh.
Two sample investigations
Scroll or swipe horizontally for Infrastructure and Content access. When the examples are focused, use the left and right arrow keys, Home or End. Each product screen can be enlarged.
Showing examples 1 of 2.
Infrastructure · Sample data
Where does Payment Gateway run?
Payment Gateway is an application in this sample. Its three servers each link to a rack (server cabinet).
Read the graph from left to right: Application → Servers → Racks
Payment Gateway runs on three servers in three racks in this sample.
Captured in VioGraph Graph Explorer · Sample infrastructure
Content access · Hyland OpenArch
Who can read confidential documents?
This saved sample links 13 documents in four confidential areas to five reader groups. Site managers have access across all four areas.
Groups → Documents → Folders
13 documents across four confidential areas, linked to five reader groups.
Hyland OpenArch · Fictional saved export
03 — Product tour
Inside VioGraph
Scroll horizontally to browse. When the gallery is focused, use the left and right arrow keys, Home or End. Each screen can be enlarged.
Showing screens 1 of 4.
Graph explorer
See how your systems connect
Follow applications to servers and racks, with the records behind each connection.
Sample infrastructure · 5 November 2026
Connectors
Bring your data together
Browse 14 catalogue entries, including Hyland OpenArch, with a readiness label for each.
Connector catalogue · 5 November 2026
Graph engines
Choose how your graph runs
Pick an administrator-enabled engine for each graph.
Installation settings · 5 November 2026
JupyterLab
Work with Python and Java
Use the Python SDK in JupyterLab, or the Java SDK in your applications.
Python example · Sample data · 5 November 2026
04 — Deployment and controls
Your environment. Your controls.
Self-hosted.
Run VioGraph and store graph data on your infrastructure, including on premises.
UI, APIs and SDKs.
Use the web UI, REST API, Python in JupyterLab, or the Java SDK.
Optional AI.
Explore without a model. Enable Ask AI under your administrator’s data policy.
Private evaluation uses Docker Compose and saved copies of source data. Production use requires validation in your environment.
Data access and optional AI
Data and access. Sign in through your identity provider. Control access with roles, workspaces and sharing; configured dataset policies filter and mask data before a person’s graph is built.
AI is a choice. Ask AI is off until an administrator configures it. The installation’s AI policy controls what the model may see. Its queries run with the caller’s graph permissions and are checked to prevent writes. If you configure an external model, data allowed by that policy may be sent to the provider.
Built to inspect. Inspect queries, returned records and the audit log, which includes an integrity check. Answers produced with a model still need review; broader model accuracy has not yet been measured.
Access rules must be configured for each dataset and graph. Recording a permission relationship in the data does not itself restrict who can query that graph.
Deployment and graph engine availability
Docker Compose is the current install path. Kubernetes, ECS and cloud deployment paths are experimental. VioGraph has been run on developer machines. A private evaluation validates the requirements for your environment before production use.
Choose one engine per graph, or build the same dataset on different engines to compare results. RyuGraph and FalkorDB are supported within the documented tests and measured limits. LadybugDB, the default, and DuckDB are experimental. DuckDB uses SQL/PGQ and does not support Ask AI. FalkorDB is optional, off by default and has a separate source-available licence.
05 — Private evaluation
Start with one question.
Agree the question, scope and success criteria before sharing data.
Agree the scope
Choose the question, data sources and access requirements together.
Walk through a sample
Review sample data, then evaluate a representative dataset.
Validate the fit
Check the answers, refresh process and operating requirements.