MentatLab
MentatLab is Mission Control for DAG-based agent workflows in private environments. It is designed around direct model/API nodes, internal services, and observable execution plans. In this stack, those model calls can target FlexInfer private endpoints while Loom governs context, MCP access, and policy boundaries.
DAG orchestration over direct model calls
MentatLab should feel closer to an operations console for workflow graphs than a chat wrapper. The graph is the contract: nodes, edges, execution plan, events, and checkpoints.
DAG-first design
Flows are explicit graphs, not hidden chat transcripts. The orchestrator can validate the graph, detect cycles, plan dependency order, and expose parallelizable branches.
Direct model/API nodes
The intended runtime shape is direct API model calls and internal services. In a private deployment, those model calls can point at FlexInfer endpoints.
Context stays governed
Loom remains responsible for MCP tool access, context memory, policy, and fleet coordination. MentatLab consumes that boundary instead of replacing it.
Operator visibility
Mission Control gives teams a place to inspect definitions, plans, run state, checkpoints, and events when workflows move from experiments to operations.
Where MentatLab fits
Mission Control interface for building, validating, monitoring, and executing DAG-based agent workflows in private environments.
Repeatable workflows where an operator needs a graph, direct model/API nodes, observable runs, and controlled private infrastructure.
A frontier coding harness replacement. MentatLab should call models and tools through explicit workflow nodes, not depend on opaque coding-agent semantics.
FlexInfer serves private models, Loom governs agent context and tools, and MentatLab composes those capabilities into workflow graphs.
Private models + policy controls + orchestration UX
Loom Core governs context routing and policy boundaries. MentatLab provides the DAG design and run-visibility layer over direct API model calls, internal services, and private FlexInfer endpoints.
Hosts private model endpoints and customization paths that workflow nodes can call through direct APIs.
Governs MCP, context, agent fleet state, sandbox execution, RBAC, and audit around workflow execution.
Provides a concrete healthcare ETL workload where graph orchestration and private AI assistance have a real data boundary.