Skip to content

Extending Maestro

This page is for developers adding new behavior to Arqon Maestro.

Main Extension Surfaces

  • speech and audio pipeline
  • command interpretation
  • editor/plugin integrations
  • workflow routing
  • model and training pipeline changes

Extension Strategy

flowchart LR
  Idea[New capability] --> Type{What kind of change?}
  Type --> Audio[Audio or VAD]
  Type --> Intent[Command or parsing]
  Type --> Integration[Editor or workflow integration]
  Type --> Models[Training or model behavior]

Practical Guidance

If you are changing audio behavior

  • validate microphone capture first
  • keep chunk lifecycle observable
  • verify partial and final endpoint behavior separately

If you are changing command behavior

  • confirm transcript shape first
  • verify parsing and selection logic independently
  • keep examples and docs in sync with the new command family

If you are adding integrations

  • define the context you need
  • define the action surface clearly
  • test the active-app and routing assumptions

If you are changing training/model behavior

  • document the data path
  • document the expected export artifacts
  • keep runtime assumptions aligned with model outputs
  1. identify the layer you are changing
  2. add instrumentation before deep edits
  3. make the smallest behavior change that proves the path
  4. update docs once behavior is confirmed
  5. preserve compatibility where possible during migrations