# FileDigest Sample Digest

## Job Summary

This sample shows the kind of Markdown artifact FileDigest produces after processing a small document packet.

- Source set: one PDF research brief, one DOCX methods appendix, one Markdown review note
- Processing backend: Modal Docling engine
- Output artifacts: `digest.md`, `manifest.json`
- Intended use: paste into ChatGPT, Claude, AI coding tools, a prompt packet, or a RAG preprocessing workflow

## Sources

### research-brief.pdf

The main source describes a document-heavy workflow where teams need to extract usable context before asking an AI system to analyze, synthesize, or compare evidence.

### methods-appendix.docx

The appendix includes supporting notes, definitions, and tables that should remain traceable to the source document.

### review-notes.md

The notes contain human observations and follow-up questions that can be placed next to the extracted document content.

## AI-Ready Context

FileDigest preserves the source boundary, creates a readable Markdown digest, and leaves an audit trail through the manifest. The goal is not to hide uncertainty. The goal is to make document preparation inspectable before the user gives the context to an LLM or downstream retrieval system.

## Suggested Prompt

Use the following prompt with this digest:

```text
Review the digest below. Extract the key claims, list the source sections that support each claim, and flag any missing evidence or conversion warnings that should be checked by a human.
```