API Output Schema
Agent execute (POST) request output schema :
{
"success": true, // Indicates that platform received and processed the request
"data":
{
"sessionId": string, // ID of the session, can be used to send another execution to the same session
"executionId": string, // Execution ID to query the result
"promptTokens": integer // Number of tokens used in prompt
}
}{
"success": false, // Indicates that platform completely failed the request
"error": string // Human-readable message. No data object is included on failures.
}Get execution logs (GET) request output schema - execution in progress :
{
"success": true,
"data":
{
"_id": string, // Internal ID
"executionStarted": datetime, // Start time
"nodeResults": [
null, // For each node
...,
null // For each node
],
"ctx":
{
"userId": string, // ID of the user calling the agent
"agentId": string, // ID of the agent
"sessionId": string, // ID of the session
"executionId": string // Same as the one in GET URL
}
}
}{
"success": false, // Indicates that platform completely failed the request
"error": string // Human-readable message. The HTTP status code reflects the error type (e.g., 400 validation, 401 auth, 403 permission, 429 rate limit, 500 server). No data object is included on failures.
}Get execution logs (GET) request output schema - execution finished :
{
"success": true,
"data":
{
"_id": string,
"executionStarted": datetime, // Worker execution start time
"executionEnded": datetime, // Worker execution end time
"nodeResults": [
inputParams?: object, // All input parameters for the worker in JSON format
{ // For each node
"ok": boolean, // True/false
"status": number, // The HTTP status code reflects the error type (e.g., 200 OK, 400 validation, 401 auth, 403 permission, 429 rate limit, 500 server).
"statusText": string, // Human-readable status text
"error": string, // (OPTIONAL: Only if OK is false) Human-Readable error text
"result": any, // Node-specific output object / array
"nodeId": number, // ID of the node
"executionStarted": datetime, // Time of node execution start
"executionFinished": datetime // Time of node execution end
},
...
],
"ctx":
{
"userId": string,
"agentId": string,
"sessionId": string,
"executionId": string
},
"time": number, // Executon time
"finalResult":
[ // This is an array because API supports agents with multiple output nodes, but the Web UI allows to create single output nodes manually only, so in most of scenarios this array contains only one element.
{
"ok": boolean,
"status": number, // The HTTP status code reflects the error type (e.g., 200 OK, 400 validation, 401 auth, 403 permission, 429 rate limit, 500 server).
"statusText": string, // Human-readable status text
"error": string, // (OPTIONAL: Only if OK is false) Human-Readable error text
"result": object, // Output of the final worker OUTPUT node outputParams in JSON format
"nodeId": number, // Node ID of the final workr OUTPUT node
"executionStarted": datetime, // ??
"executionFinished": datetime // Worker execution stop time
},
],
},
}{
"success": false, // Indicates that platform completely failed the request
"error": string // Human-readable message. The HTTP status code reflects the error type (e.g., 400 validation, 401 auth, 403 permission, 429 rate limit, 500 server). No data object is included on failures.
}Examples of node-specific RESULT outputs in different format:
// NB! Code worker nodes, Universal API nodes have output schema built into the node (that you can change during building)
// NB! LLM nodes and other nodes that come from configured providers have provider-specific and sometimes model-specific output schema
// You must always run these nodes first at least once to generate and be able to connect/see the output/output schema.
// PDF to image node
[
{
image: 'URL1',
description: '',
},
{
image: 'URL2',
description: '',
},
{
image: 'URL3',
description: '',
},
]
// Browser node output
{
error?: string | null,
sessionID: string,
requestID: string,
timestamp: Date,
downloadedFile?: string, // base64
downloadedFileName?: string // example: 7e0ed06c-01ea-4e71-a9e3-bfe0ab8517d2
meta: BrowserMeta,
generationInfo?: BrowserGenerationInfo,
value?: string,
screenshots: Record<string, string>,
}
// Fold worker
{
finalAccumulator: any, // Final accumulator result
processedItems: number, // Total items processed
totalItems: number, // Total items (processed + skipped)
skippedItems?: number, // only in safe mode
earlyTermination?: boolean,
memoryWarnings?: number, // count of memory warnings
}
// Map worker
{
results: any[],
totalItems: number, // number of inputs
successCount: number,
errorCount: number,
effectiveConcurrency: number, // max number of items that were processed in parallel
errors?: Array<{ index: number, error: string }>,
}
// Read URL
{
content: string, // Clean markdown content
title: string, // Page title
url: string, // URL that was read
}
// Storage upload
{
url: string, // Public URL of uploaded file
filename: string, // Final stored filename
secret: string, // Secret key for secure access
size: string, // File size in bytes
contentType: string, // MIME type of the file
}
// Switch worker
{
matchedCases: [{ // Array
caseIndex: number,
value: any,
workerId: string,
result: any,
success: boolean,
error?: string,
}],
defaultExecuted?: boolean,
totalCases: number,
switchValue: any,
hasAnyMatch: boolean, // Whether any case matched
executionStopped?: boolean, // Whether stopped due to break
}
// Until worker
{
finalState: any,
iterations: number,
maxIterations: number,
conditionMet: boolean,
terminationReason:
| 'condition_met'
| 'initial_condition_met'
| 'max_iterations'
| 'timeout'
| 'error',
executionTimeMs: number,
memoryWarnings?: number,
timeoutWarnings?: number,
consecutiveErrors?: number,
}
// Vector save
{
docId: string
}
// Vector search
{
success: boolean,
results: IVectorSearchResult[],
count: number,
embedder: string,
searchIndex: string,
}Updated about 1 month ago