Appearance

remora-monorepo

remora-monorepo / lib / ExecuteWorkflowOptions

Interface: ExecuteWorkflowOptions<ExecutionContext, ApprovalRequestDecisionDetails> ​

Defined in: packages/core/src/executor/executor-types.ts:57

Options for executeWorkflow.

Type Parameters ​

ExecutionContext ​

ExecutionContext extends Record<string, unknown> = Record<string, unknown>

ApprovalRequestDecisionDetails ​

ApprovalRequestDecisionDetails extends Record<string, unknown> = Record<string, unknown>

Properties ​

agent? ​

optional agent: Agent<any, any, any>

Defined in: packages/core/src/executor/executor-types.ts:72

An AI SDK Agent to use for agent-loop steps. When provided, agent-loop steps use the Agent's own tools and describe the expected output shape in the prompt. The bare model is then used to coerce the Agent's text output into the structured format. Only applies to agent-loop steps.

approvalBackoffMultiplier? ​

optional approvalBackoffMultiplier: number

Defined in: packages/core/src/executor/executor-types.ts:102

Backoff multiplier applied to the polling interval after each attempt.

See ​

DEFAULT_APPROVAL_BACKOFF_MULTIPLIER

approvalIntervalMs? ​

optional approvalIntervalMs: number

Defined in: packages/core/src/executor/executor-types.ts:100

Initial polling interval in milliseconds for approval condition checks.

See ​

DEFAULT_APPROVAL_INTERVAL_MS

approvalMaxIntervalMs? ​

optional approvalMaxIntervalMs: number

Defined in: packages/core/src/executor/executor-types.ts:104

Maximum polling interval in milliseconds. The interval will never grow beyond this.

See ​

DEFAULT_APPROVAL_MAX_INTERVAL_MS

approvalTimeoutMs? ​

optional approvalTimeoutMs: number

Defined in: packages/core/src/executor/executor-types.ts:98

Timeout in milliseconds for approval requests before treating as rejection.

See ​

DEFAULT_APPROVAL_TIMEOUT_MS

channel? ​

optional channel: WorkflowExecutionStateChannel

Defined in: packages/core/src/executor/executor-types.ts:86

Optional channel for publishing execution state snapshots. Publishes on every state transition, in addition to onStateChange.

context? ​

optional context: DurableContext

Defined in: packages/core/src/executor/executor-types.ts:88

Injectable durable execution context. Default: simple in-process implementation.

executionContext? ​

optional executionContext: ExecutionContext

Defined in: packages/core/src/executor/executor-types.ts:96

App-defined execution context passed to policy deciders (e.g. user, organization, session).

initialState? ​

optional initialState: object

Defined in: packages/core/src/executor/executor-types.ts:92

Previous execution state to resume from. When provided, completed steps are skipped and execution continues from the first incomplete step.

completedAt? ​

optional completedAt: string = "string"

durationMs? ​

optional durationMs: number = "number"

error? ​

optional error: object = errorSnapshotSchema

error.category ​

category: string = "string"

error.code ​

code: string = "string"

error.isRetryable? ​

optional isRetryable: boolean = "boolean"

error.message ​

message: string = "string"

error.statusCode? ​

optional statusCode: number = "number"

error.stepId? ​

optional stepId: string = "string"

output? ​

optional output: unknown = "unknown"

runId ​

runId: string = "string"

startedAt ​

startedAt: string = "string"

status ​

status: "pending" | "running" | "completed" | "failed" = runStatusSchema

stepRecords ​

stepRecords: object[]

workflowHash? ​

optional workflowHash: string = "string"

inputs? ​

optional inputs: Record<string, unknown>

Defined in: packages/core/src/executor/executor-types.ts:74

Input values passed to the workflow's start step. Validated against the start step's inputSchema.

limits? ​

optional limits: ExecutorLimits

Defined in: packages/core/src/executor/executor-types.ts:90

Execution limits for sleep/wait/timeout bounds.

maxRetries? ​

optional maxRetries: number

Defined in: packages/core/src/executor/executor-types.ts:76

Maximum number of retries for recoverable errors (rate limits, network errors, parse failures). Defaults to 3.

model? ​

optional model: LanguageModel

Defined in: packages/core/src/executor/executor-types.ts:67

An AI SDK LanguageModel for llm-prompt, extract-data, and agent-loop steps. Required if the workflow contains LLM steps.

onStateChange()? ​

optional onStateChange: (state, delta) => void

Defined in: packages/core/src/executor/executor-types.ts:84

Called on every state transition with the full execution state and the idempotent delta that produced it.

Parameters ​

state ​
completedAt? ​

string = "string"

durationMs? ​

number = "number"

error? ​

{ category: string; code: string; isRetryable?: boolean; message: string; statusCode?: number; stepId?: string; } = errorSnapshotSchema

error.category ​

string = "string"

error.code ​

string = "string"

error.isRetryable? ​

boolean = "boolean"

error.message ​

string = "string"

error.statusCode? ​

number = "number"

error.stepId? ​

string = "string"

output? ​

unknown = "unknown"

runId ​

string = "string"

startedAt ​

string = "string"

status ​

"pending" | "running" | "completed" | "failed" = runStatusSchema

stepRecords ​

object[] = ...

workflowHash? ​

string = "string"

delta ​

{ runId: string; startedAt: string; type: "run-started"; } | { path: ({ itemValue: unknown; iterationIndex: number; stepId: string; type: "for-each"; } | { matchedCaseIndex: number; matchedValue: unknown; stepId: string; type: "switch-case"; } | { pollAttempt: number; stepId: string; type: "wait-for-condition"; })[]; startedAt: string; stepId: string; type: "step-started"; } | { completedAt: string; durationMs: number; output: unknown; path: ({ itemValue: unknown; iterationIndex: number; stepId: string; type: "for-each"; } | { matchedCaseIndex: number; matchedValue: unknown; stepId: string; type: "switch-case"; } | { pollAttempt: number; stepId: string; type: "wait-for-condition"; })[]; resolvedInputs?: unknown; stepId: string; trace?: ({ data?: unknown; message: string; type: "log"; } | { step: unknown; type: "agent-step"; })[]; type: "step-completed"; } | { durationMs: number; error: { category: string; code: string; isRetryable?: boolean; message: string; statusCode?: number; stepId?: string; }; failedAt: string; path: ({ itemValue: unknown; iterationIndex: number; stepId: string; type: "for-each"; } | { matchedCaseIndex: number; matchedValue: unknown; stepId: string; type: "switch-case"; } | { pollAttempt: number; stepId: string; type: "wait-for-condition"; })[]; resolvedInputs?: unknown; stepId: string; type: "step-failed"; } | { path: ({ itemValue: unknown; iterationIndex: number; stepId: string; type: "for-each"; } | { matchedCaseIndex: number; matchedValue: unknown; stepId: string; type: "switch-case"; } | { pollAttempt: number; stepId: string; type: "wait-for-condition"; })[]; retry: { attempt: number; errorCode: string; errorMessage: string; failedAt: string; startedAt: string; }; stepId: string; type: "step-retry"; } | { completedAt: string; durationMs: number; output?: unknown; runId: string; type: "run-completed"; } | { durationMs: number; error: { category: string; code: string; isRetryable?: boolean; message: string; statusCode?: number; stepId?: string; }; failedAt: string; runId: string; type: "run-failed"; } | { path: ({ itemValue: unknown; iterationIndex: number; stepId: string; type: "for-each"; } | { matchedCaseIndex: number; matchedValue: unknown; stepId: string; type: "switch-case"; } | { pollAttempt: number; stepId: string; type: "wait-for-condition"; })[]; requestedAt: string; sourcePolicyId: string; stepId: string; type: "step-awaiting-approval"; } | { approvedAt: string; path: ({ itemValue: unknown; iterationIndex: number; stepId: string; type: "for-each"; } | { matchedCaseIndex: number; matchedValue: unknown; stepId: string; type: "switch-case"; } | { pollAttempt: number; stepId: string; type: "wait-for-condition"; })[]; sourcePolicyId: string; stepId: string; type: "step-approved"; } | { deniedAt: string; path: ({ itemValue: unknown; iterationIndex: number; stepId: string; type: "for-each"; } | { matchedCaseIndex: number; matchedValue: unknown; stepId: string; type: "switch-case"; } | { pollAttempt: number; stepId: string; type: "wait-for-condition"; })[]; reason?: string; sourcePolicyId: string; stepId: string; type: "step-denied"; }

Returns ​

void

onStepComplete()? ​

optional onStepComplete: (stepId, output) => void

Defined in: packages/core/src/executor/executor-types.ts:82

Called when a step completes successfully.

Parameters ​

stepId ​

string

output ​

unknown

Returns ​

void

onStepStart()? ​

optional onStepStart: (stepId, step) => void

Defined in: packages/core/src/executor/executor-types.ts:80

Called when a step begins execution.

Parameters ​

stepId ​

string

step ​

{ description: string; id: string; name: string; nextStepId?: string; params: { toolInput: {[key: string]: { type: "literal"; value: unknown; } | { expression: string; type: "jmespath"; } | { template: string; type: "template"; }; }; toolName: string; }; type: "tool-call"; } | { description: string; id: string; name: string; nextStepId?: string; params: { outputFormat: object; prompt: string; }; type: "llm-prompt"; } | { description: string; id: string; name: string; nextStepId?: string; params: { outputFormat: object; sourceData: { type: "literal"; value: unknown; } | { expression: string; type: "jmespath"; } | { template: string; type: "template"; }; }; type: "extract-data"; } | { description: string; id: string; name: string; nextStepId?: string; params: { cases: object[]; switchOn: { type: "literal"; value: unknown; } | { expression: string; type: "jmespath"; } | { template: string; type: "template"; }; }; type: "switch-case"; } | { description: string; id: string; name: string; nextStepId?: string; params: { itemName: string; loopBodyStepId: string; target: { type: "literal"; value: unknown; } | { expression: string; type: "jmespath"; } | { template: string; type: "template"; }; }; type: "for-each"; } | { description: string; id: string; name: string; nextStepId?: string; params: { durationMs: { type: "literal"; value: unknown; } | { expression: string; type: "jmespath"; } | { template: string; type: "template"; }; }; type: "sleep"; } | { description: string; id: string; name: string; nextStepId?: string; params: { backoffMultiplier?: { type: "literal"; value: unknown; } | { expression: string; type: "jmespath"; } | { template: string; type: "template"; }; condition: { type: "literal"; value: unknown; } | { expression: string; type: "jmespath"; } | { template: string; type: "template"; }; conditionStepId: string; intervalMs?: { type: "literal"; value: unknown; } | { expression: string; type: "jmespath"; } | { template: string; type: "template"; }; maxAttempts?: { type: "literal"; value: unknown; } | { expression: string; type: "jmespath"; } | { template: string; type: "template"; }; timeoutMs?: { type: "literal"; value: unknown; } | { expression: string; type: "jmespath"; } | { template: string; type: "template"; }; }; type: "wait-for-condition"; } | { description: string; id: string; name: string; nextStepId?: string; params: { instructions: string; maxSteps?: { type: "literal"; value: unknown; } | { expression: string; type: "jmespath"; } | { template: string; type: "template"; }; outputFormat: object; tools: string[]; }; type: "agent-loop"; } | { description: string; id: string; name: string; nextStepId?: string; type: "start"; } | { description: string; id: string; name: string; nextStepId?: string; params?: { output: { type: "literal"; value: unknown; } | { expression: string; type: "jmespath"; } | { template: string; type: "template"; }; }; type: "end"; }

Returns ​

void

policies? ​

optional policies: Policy<ExecutionContext, ApprovalRequestDecisionDetails>[]

Defined in: packages/core/src/executor/executor-types.ts:94

Authorization policies evaluated in order before each tool-call step. When absent, all steps execute freely.

retryDelayMs? ​

optional retryDelayMs: number

Defined in: packages/core/src/executor/executor-types.ts:78

Base delay in milliseconds for exponential backoff between retries. Defaults to 1000.

tools ​

tools: ToolSet

Defined in: packages/core/src/executor/executor-types.ts:65

Tool definitions. Every tool referenced by a tool-call step must be present with an execute function.