dMetrics

Founded 2012

Convert unstructured enterprise documents into verifiable knowledge graphs with citations

Pricing model

Demo

You can use dMetrics to turn siloed operational data and technical documents into verified knowledge graphs. The system links every generated insight directly to source evidence and tradecraft standards, avoiding hallucinations from ungrounded language models. You can encode human domain expertise into interactive ontologies without engineering support, evaluate risk across deployed machine learning models, and control data access down to individual graph nodes. It also supports secure deployments across sensitive defense networks and commercial clouds.

What dMetrics does

Generate cited answers from unstructured document repositoriesMap custom ontologies and relationship graphs from raw textAudit language models for mission-critical risk and biasEnforce node-level and edge-level security permissions on graphsVerify factual claims against primary source evidenceExtract entities and resolve identities across heterogeneous datasetsTrack science and technology spending against milestonesTrace every analytical answer directly back to source documents and citationsBuild and edit custom domain ontologies without relying on software engineersEnforce fine-grained access controls at the individual graph node and edge levelAutomate risk and performance assessments for deployed machine learning modelsExtract entities and map relationships across disconnected, heterogeneous filesDeploy directly in air-gapped environments or FedRAMP and DoD IL-4 clouds