FlowiseAI/Flowise · error · Error

Invalid JSON object. Error: ${error}

Error message

Invalid JSON object. Error: ${error}

What it means

ConditionAgent.parseJsonMarkdown locates a JSON block inside the LLM's response (handling ```json fences and bare braces), slices it out, and calls JSON.parse. If parse fails on the extracted substring, it throws 'Invalid JSON object. Error: <SyntaxError>'. This means a JSON-looking region was found but is syntactically malformed.

Source

Thrown at packages/components/nodes/agentflow/ConditionAgent/ConditionAgent.ts:242

        // Find end of JSON
        if (startIndex !== -1) {
            for (const e of ends) {
                endIndex = jsonString.lastIndexOf(e, jsonString.length)
                if (endIndex !== -1) {
                    if (jsonString[endIndex] === '}') {
                        endIndex += 1
                    }
                    break
                }
            }
        }

        if (startIndex !== -1 && endIndex !== -1 && startIndex < endIndex) {
            const extractedContent = jsonString.slice(startIndex, endIndex).trim()
            try {
                return JSON.parse(extractedContent)
            } catch (error) {
                throw new Error(`Invalid JSON object. Error: ${error}`)
            }
        }

        throw new Error('Could not find JSON block in the output.')
    }

    async run(nodeData: INodeData, question: string, options: ICommonObject): Promise<any> {
        let llmIds: ICommonObject | undefined
        let analyticHandlers = options.analyticHandlers as AnalyticHandler

        try {
            const abortController = options.abortController as AbortController

            // Extract input parameters
            const model = nodeData.inputs?.conditionAgentModel as string
            const modelConfig = nodeData.inputs?.conditionAgentModelConfig as ICommonObject
            if (!model) {
                throw new Error('Model is required')

View on GitHub (pinned to abe4a8601a)

Solutions

1. Use a stronger model or one with native JSON/structured-output mode for the Condition Agent.
2. Lower temperature and tighten the system prompt / few-shot examples.
3. Increase max_tokens so the JSON object is not truncated.
4. Pre-validate the LLM string with a tolerant JSON parser (jsonrepair) before parseJsonMarkdown, or sanitize common issues (trailing commas, single quotes).

Example fix

// before
const parsed = parseJsonMarkdown(responseContent) // throws on trailing comma

// after — sanitize then parse
import { jsonrepair } from 'jsonrepair'
const repaired = jsonrepair(extractedContent)
const parsed = JSON.parse(repaired)

import { jsonrepair } from 'jsonrepair'
function safeParseJsonMarkdown(extracted) {
  let repaired
  try { repaired = jsonrepair(extracted) }
  catch { throw new Error('LLM output is not repairable JSON') }
  return JSON.parse(repaired)
}

function looksLikeJsonObject(s) {
  const t = (s ?? '').trim(); return t.startsWith('{') && t.endsWith('}') }

try { parsed = parseJsonMarkdown(responseContent) }
catch (e) {
  if (/Invalid JSON object/.test(e.message)) {
    // retry with a stricter prompt or a stronger model
  } else throw e
}

When it happens

Trigger: LLM returns a fenced block that looks like JSON but contains trailing commas, single quotes, unquoted keys, comments, or truncated content; model output with mixed prose inside the braces; partial token streaming that cut off mid-object.

Common situations: Weaker/smaller model not following JSON instructions; temperature too high producing valid-ish but invalid JSON; max_tokens cutting the response mid-object; prompt change that confuses the model's output format.

Understand the failure class

• Parsing and encoding errors: unexpected token, malformed input — why parsers reject input and how to find the real culprit.

Related errors

• Could not find JSON block in the output.
• LLM response is missing the "output" key or it is not a stri
• Failed to parse a valid scenario from the LLM's response. Pl
• chatflowId must be a valid array
• Model is required

AI-assisted analysis of FlowiseAI/Flowise@abe4a8601a (2026-08-12). Data as JSON: /api/errors/e95672a6133b6741. Report an issue: GitHub.