Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.mindlab.io/llms.txt

Use this file to discover all available pages before exploring further.

The core philosophy of the MindLab platform is “Configuration-first.” We believe that for AI to be a trusted, enterprise-grade asset, its behavior must be deterministic, repeatable, and auditable. This cannot be achieved through prompting alone. This is why we created the CADANCE™ Spec, a single, versioned specification that declaratively defines the entire lifecycle of an AI-driven workflow. It is a human-readable, machine-executable contract that ensures every action taken by the system is traceable, reproducible, and aligned with your explicit intent.

​
The 7 Components of the CADANCE™ Spec

CADANCE™ is an acronym for the seven core components of the specification:
Defines the high-level objectives, success criteria, and operational budgets (cost, latency, SLOs) for the workflow.
Specifies the roles, capabilities, constraints, and escalation paths for the specialist agents involved in the workflow.
Declares the knowledge sources, memory scopes, data contracts, and citation requirements for the workflow.
Defines the output templates, evaluation rubrics, and schemas that the agents must adhere to.
Specifies the policies, governance rules, approval requirements, and human-in-the-loop (HITL) checkpoints for the workflow.
Defines the state graphs, retry/timeout logic, gating conditions, and deterministic seeds for the workflow’s execution.
Specifies the runbooks, handoffs, logging levels, artifact retention policies, and audit requirements for the workflow.