Production Demo Run
A reproducible FileDigest run using public input files, the live app, production processing, and raw downloadable outputs.
This page documents the public demo run behind FileDigest Examples. It uses public source files instead of fictional packet names or private customer material. The packet was first published in April 2026 and is regenerated through the production pipeline as the engine evolves; the artifacts and figures below reflect the current run.
Production domain: filedigest.dev
Job ID: df56be0156354d259b5b63b4e08dabd4
How the packet was produced
- Seven public/permissively licensed files were collected and archived.
- The files were uploaded through the app into private storage.
- The job was processed by the production Modal Docling engine.
- The generated
digest.mdandmanifest.jsonwere downloaded from the job detail page. - The production job was deleted after the public artifact copies were saved.
Download the packet
Result
| Check | Result |
|---|---|
| Job status | SUCCEEDED |
| Parsed files | 7 of 7 |
| Failed files | 0 |
| Warnings | None |
| Output tokens | 24,017 |
| Worker time | 21.3 seconds |
| Processing mode | Fast text extraction, accurate tables |
| OCR | Automatic (the scanned page was detected and retried with OCR) |
Input files
| Source | File type | Public artifact |
|---|---|---|
| NIST AI Risk Management Framework (NIST AI 100-1) | nist-ai-risk-management-framework.pdf | |
| Image-only scanned page (auto-OCR showcase) | scanned-field-log.pdf | |
| NASA Earth Lithograph | Earth_Lithograph.pdf | |
| file-format-commons sample | DOCX | ffc.docx |
| file-format-commons sample | PPTX | ffc.pptx |
| MDN beginner HTML sample | HTML | mdn-beginner-html-index.html |
| Public README template | Markdown | good-readme-template.md |
What to inspect
Open digest.md to judge the human-readable output: source boundaries, processing settings, extracted text, and the source index.
Open manifest.json to inspect structured metadata for automation, audit, ingestion, or repeatability.
Open provenance.json for the publication trail: source URLs, license notes, SHA-256 hashes of every input and output, and the job ID.
Limitations
This is one public demo packet, not a universal benchmark. It does not prove how every scanned, damaged, encrypted, image-heavy, or unusually formatted file will parse. It does show the exact output contract produced by the live app on a mixed public packet.
Download the example outputs, read security details, or start your own packet.