FlowiseAI/Flowise · error · Error

Failed to process paper "${result.title}": ${errorMessage}

Error message

Failed to process paper "${result.title}": ${errorMessage}

What it means

Thrown in the per-paper processing loop of ArxivTool._call when downloadAndExtractPdf or document construction throws AND this.continueOnFailure is false. The title and the inner errorMessage are interpolated. The tool supports a graceful mode (continueOnFailure=true) that downgrades the same failure to a fallback summary, so this hard error is a deliberate 'strict mode' behavior.

Source

Thrown at packages/components/nodes/tools/Arxiv/core.ts:246

                        // Download and extract PDF content
                        const fullText = await this.downloadAndExtractPdf(result.id)

                        const publishedDate = result.published ? new Date(result.published).toISOString().split('T')[0] : 'Unknown'

                        // Format with metadata and full content
                        const docContent = `Published: ${publishedDate}\nTitle: ${result.title}\nAuthors: ${result.authors.join(
                            ', '
                        )}\nSummary: ${result.summary}\n\nFull Content:\n${fullText}`

                        const truncatedContent = this.docContentCharsMax ? docContent.substring(0, this.docContentCharsMax) : docContent

                        docs.push(truncatedContent)
                    } catch (error) {
                        const errorMessage = error instanceof Error ? error.message : 'Unknown error'
                        console.error(`Error processing paper ${result.title}:`, errorMessage)

                        if (!this.continueOnFailure) {
                            throw new Error(`Failed to process paper "${result.title}": ${errorMessage}`)
                        } else {
                            // Add error notice and continue with summary only
                            const publishedDate = result.published ? new Date(result.published).toISOString().split('T')[0] : 'Unknown'
                            const fallbackContent = `Published: ${publishedDate}\nTitle: ${result.title}\nAuthors: ${result.authors.join(
                                ', '
                            )}\nSummary: ${result.summary}\n\n[ERROR: Could not load full content - ${errorMessage}]`
                            docs.push(fallbackContent)
                        }
                    }
                }

                return docs.join('\n\n---\n\n')
            }
        } catch (error) {
            const errorMessage = error instanceof Error ? error.message : 'Unknown error'
            console.error('Arxiv search error:', errorMessage)
            throw new Error(`Failed to search Arxiv: ${errorMessage}`)
        }

View on GitHub (pinned to abe4a8601a)

Solutions

1. Set continueOnFailure=true on the tool so a single bad paper does not abort the whole batch — the tool will return a fallback summary for that paper.
2. Identify the failing paper by its title in the error and either exclude it or fix its source.
3. If full text is not required, set loadFullContent=false to skip PDF download entirely.

Example fix

// before
const tool = new ArxivTool({ continueOnFailure: false, loadFullContent: true })
// after
const tool = new ArxivTool({ continueOnFailure: true, loadFullContent: true })
// failing papers now degrade to a summary with [ERROR: ...] instead of aborting

function shouldContinueOnFailure(cfg: { loadFullContent?: boolean; continueOnFailure?: boolean }) {
  // Recommend graceful mode whenever full content is enabled in batch runs
  return cfg.loadFullContent ? Boolean(cfg.continueOnFailure) : true
}

try {
  docs.push(await processPaper(result))
} catch (e) {
  if (!continueOnFailure) throw e
  docs.push(fallbackSummary(result, e))
}

When it happens

Trigger: A PDF download fails (see error 349) or PDFLoader fails to parse the PDF while continueOnFailure is false; the loop hits any error during full-content ingestion and is configured to stop on first failure.

Common situations: Default configuration that requires every paper's full text; ingestion of a corpus containing one withdrawn or corrupt PDF; PDFjs version incompatibility causing parse errors.

Related errors

• Failed to download PDF: ${response.status} ${response.status
• Failed to search Arxiv: ${errorMessage}
• Model is required
• Invalid Flow State
• Tool not selected

AI-assisted analysis of FlowiseAI/Flowise@abe4a8601a (2026-08-12). Data as JSON: /api/errors/5a46846a63d54e42. Report an issue: GitHub.