{"observation":{"id":"e260a099-6895-499f-ab5d-82e11a0cf513","tool":"upstage-ai","tool_name":"Upstage AI","criterion":"advanced-features","criterion_name":"Advanced Features","criterion_definition":"Provides separate table/chart extraction and flags low-confidence OCR or ambiguous regions.","criterion_evidence_type":"transformation","criterion_rank_role":"context","criterion_rank_role_reason":"Separate extraction modes and OCR confidence flags are useful workflow features, but they do not by themselves determine whether the Markdown conversion is good. (3 of 3 judges)","scenario":"hybrid-earnings-report","scenario_name":"Hybrid Earnings Report","group_tag":"hybrid-earnings-report","scenario_description":"A long, real-world hybrid annual report used to benchmark end-to-end PDF-to-markdown conversion. It combines native text, financial tables, charts/graphics, and scanned signature/stamp regions, stressing preservation of document structure, reading order, and embedded visual content across a multi-section report.","modality":"pdf","input_text":null,"input_artifact_refs":[{"alt":null,"url":"https://cdn.futuresmart.ai/public/aidemos/f63a44b441704753b65d1b89dc9b4bbb.pdf?v=1","role":"input","filename":"Target-2015-Annual-Report.pdf"}],"stresses":["Native digital text extraction","Complex financial table preservation","Chart and graphic retention","Image-based signatures and stamps","Reading order across a long multi-section report","Markdown quality and consistency"],"verdict":"worked","score":null,"score_total":null,"note":"Separately extracts chart content into textual summaries and value tables, including a waterfall chart and bar charts with year/value pairs and growth statistics.","evidence_state":"verified","source":null,"artifacts":[{"url":"https://cdn.futuresmart.ai/public/aidemos/bb341a5637d24cd7b0d5b4c5bb9effb5.png?v=1","role":"input","alt":null},{"url":"https://cdn.futuresmart.ai/public/aidemos/db32f9b4633642b88ab3a68a40562019.png?v=1","role":"output","alt":null},{"url":"https://cdn.futuresmart.ai/public/aidemos/7e9c4cab28e4455f8df43bb8f0d3ee1e.png?v=1","role":"output","alt":null},{"url":"https://cdn.futuresmart.ai/public/aidemos/44d360ec006d489ea0417444c062df4f.png?v=1","role":"output","alt":null}],"run_id":"6e3160de-fe46-4b45-b071-72560b5c5d0e","study_title":"Convert a Complex PDF into Clean Markdown with an API","study_kind":"generation","research_task":"86b9h7t37","tested_at":null,"completeness":"input-and-output","input":{"state":"files","text":null,"files":[{"url":"https://cdn.futuresmart.ai/public/aidemos/f63a44b441704753b65d1b89dc9b4bbb.pdf?v=1","filename":"Target-2015-Annual-Report.pdf","alt":"Hybrid Earnings Report","role":"input"}],"modality":"pdf","stresses":["Native digital text extraction","Complex financial table preservation","Chart and graphic retention","Image-based signatures and stamps","Reading order across a long multi-section report","Markdown quality and consistency"]},"tool_page_slug":"upstage-ai","tool_url":"https://aidemos.com/tools/upstage-ai","permalink":"https://aidemos.com/evidence/e260a099-6895-499f-ab5d-82e11a0cf513","api_url":"https://ai.aidemos.com/v1/observations/e260a099-6895-499f-ab5d-82e11a0cf513"},"peers":[{"id":"51f08c65-3c19-4c23-8666-87b5d8a58831","tool":"landing-ai","tool_name":"Landing AI","verdict":"worked","score":null,"score_total":null,"note":"Emits semantic attestation blocks for handwritten-signature regions and a blurry signature stamp, distinguishing the signer/company name and the signature legibility instead of leaving those regions unannotated.","artifact_count":5,"thumbnail":"https://cdn.futuresmart.ai/public/aidemos/1f45fe222ffc4719ad527014bd6fabbf.png?v=1","evidence_url":"https://aidemos.com/evidence/51f08c65-3c19-4c23-8666-87b5d8a58831"},{"id":"5b164bff-8ffe-46e2-b493-883b4ed9fed4","tool":"reducto","tool_name":"Reducto","verdict":"worked","score":null,"score_total":null,"note":"The separate schema-driven extract.run endpoint cleanly recovers both a financial table and a donut chart: all requested table rows and all five donut percentages come back exactly.","artifact_count":4,"thumbnail":"https://cdn.futuresmart.ai/public/aidemos/27079c195497426b998691e2f7749eea.png?v=1","evidence_url":"https://aidemos.com/evidence/5b164bff-8ffe-46e2-b493-883b4ed9fed4"},{"id":"324c753c-f3ab-4165-8bd1-eda961278d78","tool":"tensorlake","tool_name":"Tensorlake","verdict":"worked","score":null,"score_total":null,"note":"Provides separate chart extraction for the SG&A waterfall, outputting chart metadata with a title, axis labels, 10 category labels, and a numeric series instead of only inline prose.","artifact_count":2,"thumbnail":"https://cdn.futuresmart.ai/public/aidemos/bb341a5637d24cd7b0d5b4c5bb9effb5.png?v=1","evidence_url":"https://aidemos.com/evidence/324c753c-f3ab-4165-8bd1-eda961278d78"}],"other_criteria":[{"id":"4f68e405-a4c7-4191-b4bc-3c3a430b14b4","criterion":"complex-document-handling","criterion_name":"Complex Document Handling","rank_role":"decisive","verdict":"mixed","score":null,"score_total":null,"note":"Handles an 84-page hybrid report end to end, but quality degrades on signatures and multicolumn pages rather than staying uniform across the document.","artifact_count":4,"evidence_url":"https://aidemos.com/evidence/4f68e405-a4c7-4191-b4bc-3c3a430b14b4"},{"id":"07e93c54-d160-40c0-b3ba-519704752930","criterion":"reading-order-structure","criterion_name":"Reading Order & Structure","rank_role":"decisive","verdict":"failed","score":null,"score_total":null,"note":"Breaks the two-column reading order on the annual-report page, so headings and body text no longer follow the source navigation cleanly.","artifact_count":2,"evidence_url":"https://aidemos.com/evidence/07e93c54-d160-40c0-b3ba-519704752930"},{"id":"eadd62f1-5229-47c5-9d69-024cbf87e51b","criterion":"table-preservation","criterion_name":"Table Preservation","rank_role":"decisive","verdict":"mixed","score":null,"score_total":null,"note":"Reconstructs the 2011–2015 financial summary table with rows and columns intact, but misses a small number of currency symbols in the extracted values.","artifact_count":2,"evidence_url":"https://aidemos.com/evidence/eadd62f1-5229-47c5-9d69-024cbf87e51b"},{"id":"417a2325-3f7b-4e6f-a06f-e7653967e43e","criterion":"visual-content-retention","criterion_name":"Visual Content Retention","rank_role":"decisive","verdict":"failed","score":null,"score_total":null,"note":"Does not retain handwritten signature content as visual output on the signatures page, leaving only partial surrounding legal text and no clearly identifiable signatures.","artifact_count":2,"evidence_url":"https://aidemos.com/evidence/417a2325-3f7b-4e6f-a06f-e7653967e43e"}],"appears_in":[{"page_type":"ranking","slug":"pdf-to-markdown-apis","title":"Best AI Tools to Convert Complex PDFs into Clean Markdown with an API","url":"https://aidemos.com/best/pdf-to-markdown-apis","binding":"run"}],"same_scenario":[{"id":"51f08c65-3c19-4c23-8666-87b5d8a58831","tool":"landing-ai","tool_name":"Landing AI","verdict":"worked","score":null,"score_total":null,"note":"Emits semantic attestation blocks for handwritten-signature regions and a blurry signature stamp, distinguishing the signer/company name and the signature legibility instead of leaving those regions unannotated."},{"id":"5b164bff-8ffe-46e2-b493-883b4ed9fed4","tool":"reducto","tool_name":"Reducto","verdict":"worked","score":null,"score_total":null,"note":"The separate schema-driven extract.run endpoint cleanly recovers both a financial table and a donut chart: all requested table rows and all five donut percentages come back exactly."},{"id":"324c753c-f3ab-4165-8bd1-eda961278d78","tool":"tensorlake","tool_name":"Tensorlake","verdict":"worked","score":null,"score_total":null,"note":"Provides separate chart extraction for the SG&A waterfall, outputting chart metadata with a title, axis labels, 10 category labels, and a numeric series instead of only inline prose."}]}