Intelligence
The LLM substrate — labs and their brain, ears, and mouth.
Intelligence is the language-model layer: a `lab` is the workspace, and its compound pieces are the modalities — `brain` for reasoning over a chosen model, `ears` for speech-to-text, `mouth` for text-to-speech. Reach for intelligence when the work is open-ended language: summarize, classify, converse, transcribe, speak. Model choice and provider routing happen behind one interface, so you describe the intent, not the vendor.
Capabilities
| "Element" | "access" | "activity" | "attachments" | "batch_stats" | "compose" | "context" | "create" | "delete" | "disable" | "enable" | "export_bundle" | "get" | "import_bundle" | "intention" | "promote" | "readme" | "readme_update" | "remove-modifier" | "restore" | "schema" | "source" | "source_branches" | "source_promote" | "source_repair" | "source_status" | "source_validate" | "stats" | "tree" | "update" | "update_meta" | "version" |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| "Brain" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" |
| "Ears" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" |
| "Lab" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" |
| "Mouth" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" | "✓" |