API Reference
This section provides a complete reference for the R-Think public API. All interfaces, methods, and types are documented below.
Runtime
The Runtime class is the primary entry point for creating and executing cognitive missions.
| Method | Path | Description |
|---|---|---|
| Static | Runtime.create(options) | Create a new mission instance with protocol, rules, and artifact configuration. |
| Instance | runtime.execute(input) | Execute the mission with the given input. Returns validated output with trace and artifacts. |
| Instance | runtime.validate(input) | Validate input against mission rules without executing the protocol. |
| Instance | runtime.trace() | Get the execution trace of the last mission run. |
| Instance | runtime.stop() | Stop a running mission execution. |
Runtime.create Options
| Parameter | Type | Required | Description |
|---|---|---|---|
| id | string | Required | Unique identifier for the mission instance. |
| protocol | ProtocolDefinition | Required | The protocol definition to execute. |
| rules | Rule[] | Optional | Array of validation rules to apply at runtime. |
| artifacts | ArtifactConfig[] | Optional | Array of artifact capture configurations. |
| adaptiveDepth | AdaptiveDepth | Optional | Mission risk level L0 through L3. Defaults to L1. |
| retry | RetryConfig | Optional | Retry configuration for failed steps. |
Response Schema
The execute method returns a MissionResult object with the following structure:
Response
{
"output": { "result": "decision-ready" },
"validated": true,
"trace": {
"id": "trace-001",
"missionId": "mission-001",
"steps": [
{ "name": "observe", "passed": true, "durationMs": 12 }
],
"startedAt": "2026-07-22T10:00:00Z",
"completedAt": "2026-07-22T10:00:02Z"
},
"artifacts": { "evidence": [], "metrics": {} },
"evidenceGraph": { "nodes": [], "edges": [] }
}