QuantumBPM

QuantumBPM gives agentic and AI workflows the guardrails they need: defined process paths, deterministic rules, human approval steps, and a complete audit trail. It combines a BPMN 2.0 process engine and a DMN 1.5 decision engine with level 3 conformance in a single product. Teams can visually model the workflows their agents run inside and the decision tables that govern what happens next, using the same FEEL expression language across both. AI agents and services plug in as external workers in any language.

Every process instance and decision evaluation is recorded with its full inputs, outputs, and state. When an automated workflow goes wrong, you can scrub through its execution with a replay slider, resolve incidents and send messages or signals from the UI, and test a tightened rule against past executions before promoting it.

The runtime is written entirely in Go. It ships as a single binary with sub-millisecond decision evaluation and no JVM warm-up, and its DMN engine passes 99.32% of the community conformance test kit. You can integrate through a REST API or SDKs for Go, Java, TypeScript, and Python. It runs as managed SaaS or self-hosted.

For teams evaluating Camunda, it's a lighter alternative with broader BPMN and DMN specification coverage and user experience.