FlowiseAI/Flowise · error · Error

Error submitting tool outputs. Thread ID: ${threadId}. Run I

Error message

Error submitting tool outputs. Thread ID: ${threadId}. Run ID: ${runThreadId}

What it means

After tools execute, their outputs are submitted back to the OpenAI run via handleToolSubmission. If that submission call fails (network, OpenAI 4xx/5xx, run already expired/cancelled), the code cancels the run, reports onLLMError/onChainError, and throws this message. The original submission error is logged but not in the thrown message.

Source

Thrown at packages/components/nodes/agents/OpenAIAssistant/OpenAIAssistant.ts:622

                                    chatId,
                                    options,
                                    input,
                                    usedTools,
                                    text,
                                    isStreamingStarted
                                })
                                text = result.text
                                isStreamingStarted = result.isStreamingStarted
                            } catch (error) {
                                console.error('Error submitting tool outputs:', error)
                                await openai.beta.threads.runs.cancel(runThreadId, { thread_id: threadId })

                                const errMsg = `Error submitting tool outputs. Thread ID: ${threadId}. Run ID: ${runThreadId}`

                                await analyticHandlers.onLLMError(llmIds, errMsg)
                                await analyticHandlers.onChainError(parentIds, errMsg, true)

                                throw new Error(errMsg)
                            }
                        }
                    }
                }

                // List messages
                const messages = await openai.beta.threads.messages.list(threadId)
                const messageData = messages.data ?? []
                const assistantMessages = messageData.filter((msg) => msg.role === 'assistant')
                if (!assistantMessages.length) return ''

                // Remove images from the logging text
                let llmOutput = text.replace(imageRegex, '')
                llmOutput = llmOutput.replace('<br/>', '')

                await analyticHandlers.onLLMEnd(llmIds, llmOutput)
                await analyticHandlers.onChainEnd(parentIds, messageData, true)
                return {

View on GitHub (pinned to abe4a8601a)

Solutions

Read server logs for `Error submitting tool outputs:` which prints the underlying error.
Make tools faster so submission happens before the run expires (~10 min limit).
For 429s, reduce concurrency or add backoff; for 5xx, retry the run.
Ensure the thread/run is not being cancelled elsewhere concurrently.
Patch the message to include the cause for easier triage.

Example fix

// before
                                const errMsg = `Error submitting tool outputs. Thread ID: ${threadId}. Run ID: ${runThreadId}`
// after
                                const errMsg = `Error submitting tool outputs. Thread ID: ${threadId}. Run ID: ${runThreadId}. Cause: ${error instanceof Error ? error.message : String(error)}`

Defensive patterns

Strategy: retry

Try / catch

for (let attempt = 0; attempt < 3; attempt++) {
  try {
    return await assistantNode.run(nodeData, input, options)
  } catch (e) {
    if ((e as Error).message.startsWith('Error submitting tool outputs.') && attempt < 2) {
      await new Promise((r) => setTimeout(r, 1000 * 2 ** attempt)); continue
    }
    throw e
  }
}

Prevention

Keep tool execution well under the OpenAI run lifetime (~10 min).
Reduce concurrency to avoid 429s on submission.
Avoid concurrent cancellation of the same thread/run.
Use streaming submission to surface failures earlier.

When it happens

Trigger: openai.beta.threads.runs.submitToolOutputs (or the stream) returns an error; the runThreadId is no longer active (expired, already cancelled); rate limit (429); network failure mid-submission.

Common situations: Tool took too long so the run expired before submission; OpenAI rate limit; transient network blip; concurrent modification/cancellation of the thread.

Related errors

AI-assisted analysis of FlowiseAI/Flowise@abe4a8601a (2026-08-12). Data as JSON: /api/errors/740f8c333b864969. Report an issue: GitHub.