Evidence · first-party tested

Breaks down on complex scanned tables by disrupting column boundaries and value alignment, which makes the reconstructed table much less faithful to the source structure.

✗ Failed🧾 artifact-verifiedTested Jun 20, 2026Mistral AI
Output — unretouched
Stand structure table comparing before and after cutting.
Stand structure table comparing before and after cutting.
Output — unretouched
Mistral AI — mistral_scannedpdf_parsed_table.png
Mistral AI — mistral_scannedpdf_parsed_table.png
Output — unretouched
Best AI Tools to Convert Complex PDFs into Clean Markdown via API — scanned_pdf_table_with_text_between_columns.png
Best AI Tools to Convert Complex PDFs into Clean Markdown via API — scanned_pdf_table_with_text_between_columns.png
Output — unretouched
Mistral AI — mistralai_scannedpdf_multicolumn_table_with_intervening_text_parsed.png
Mistral AI — mistralai_scannedpdf_multicolumn_table_with_intervening_text_parsed.png
The observation

Breaks down on complex scanned tables by disrupting column boundaries and value alignment, which makes the reconstructed table much less faithful to the source structure.

Criterion: Table Preservation · Scenario: Scanned Research Paper (group: scanned-research-paper)
Query this
get_evidence({
  tool: "mistral-ai",
  scenario: "scanned-research-paper"
})
MCP · mcp.aidemos.com/api/mcp
Free with attribution.
Real outputs, no retouching · every cell queryable via API & MCP · aidemos.com