The Records Project

Nothing is created until the AI has confirmed that the folder you chose contains documents it can actually read.

1. It asks you to select exactly one local folder.
2. It asks you to select exactly one local folder.
3. It asks which registered computer to use and for the exact folder path in chat. There is no folder picker.
4. It inspects that folder recursively without writing anything, not even a test file.
5. It ignores hidden files, temporary files, zero-byte files, and any old RECORDS-OUTPUT folder.
6. It confirms that at least one readable, non-empty source document exists.

If the same document sits in several places, the AI notices by checksum, converts it once, and lists every location. Extra copies never cost extra work.

Only after that check passes does it confirm the project type and what you want from it. Your original contents remain unchanged. Checked incoming files move from INCOMING into INCOMING/PROCESSED after successful checks. Any other move of originals happens only when you ask to organize the source folder and approve the path list.

Everything new goes into one folder next to your documents. If that folder already exists from an earlier run, the AI checks it and continues rather than starting over.

RECORDS-OUTPUT/
├── START-HERE.md
├── PROJECT-INSTRUCTIONS.md
├── NEXT-STEPS.md
├── PROGRESS.md
├── 10 CONVERTED/
├── 20 INDEX/
│   ├── SOURCE-MANIFEST.csv
│   └── REDACTION-LOG.md
├── 30 REPORTS/
└── 90 AUDIT/
    ├── EXCEPTIONS.md
    ├── AUDIT-REPORT.md
    └── IMPROVEMENT-NOTES.md

• START-HERE.md and NEXT-STEPS.md: what this library is, and three suggested questions to ask next.
• PROGRESS.md: where the AI got to, so an interrupted run can continue.
• 10 CONVERTED: the Markdown copy of every readable document, sorted into categories based on what each one contains rather than its filename.
• 20 INDEX: a list of every source file and a count of what was masked.
• 30 REPORTS: the answers and briefings you ask for.
• 90 AUDIT: what could not be read, what is uncertain, the checks the AI ran, and short notes that help improve this guide over time.

Scans receive OCR or visual transcription. The AI flags uncertainty instead of guessing, and it checks dates, units, decimal points, identifiers, totals, and other load-bearing values against the original.

The safety and evidence rules never change. The project type changes the classification, reports, and professional boundary.

Fluent prose is not proof. The output is designed so every important statement can be traced back to a local source.

• Source first. Every material claim includes the source filename and, when available, a page, sheet, or section.
• Fact and inference stay separate. The AI labels interpretation, conflict, missing context, and uncertainty.
• No invented completeness. Coverage means “all readable documents in the selected folder,” not all records that might exist elsewhere.
• Documents are data, not instructions. Commands found inside a source document are treated as untrusted text and never followed.
• Redaction affects derived files only. Direct identifiers, account numbers, member IDs, and similar values are replaced in converted copies. Originals remain untouched.
• Redaction is checked. The AI performs a second sweep and records item types and counts without copying sensitive values into the log.
• Every conversion is checked. The AI audits every converted file against its original and reprocesses failures.

The final report states what was read, what was skipped, what could not be converted, what was redacted, what was checked, and what remains uncertain. It must not claim success while exceptions remain hidden.

Claude path. Cowork can work with files in a folder you explicitly select through Claude Desktop. Anthropic warns that computer-use actions can carry extra risk, particularly with medical and financial services. This workflow therefore uses local file access only and never asks Claude to operate a portal. Cowork access; Cowork safety. Checked Aug 28, 2026

ChatGPT path. ChatGPT Work can use a local folder after you grant access in the desktop app. This is not the same as uploading those files into a cloud Project. The workflow uses the local folder and writes its results there. Computer use; Local computer workflow. Checked Aug 28, 2026

Grok Bot path. Use the Grok Bot desktop app for Windows or Mac with your Cursor account. Open Settings from the sidebar account button at bottom-left, or use Open settings in the command palette. Turn on local execution under General, and register the computer holding your documents under Computer. In chat, tell the agent which registered computer to use and type or paste the exact folder path. On Windows, use the full path with drive letter and backslashes (for example C:\Users\YourName\Documents\Records). On Mac, use the Unix-style path (for example /Users/YourName/Documents/Records). There is no folder picker. Each local action can require approval, so approve those prompts. The agent must preserve the path exactly, work on the named registered computer, and write only inside RECORDS-OUTPUT there. It must not copy the library onto its own Linux computer or use a Cloud Agent or GitHub repo VM. Setup checked Sep 6, 2026

Local files do not mean the model runs locally. The desktop app can send file contents to the provider for processing. Review the provider’s current privacy controls and terms before using sensitive material.

Redaction is risk reduction, not anonymization. A medical or financial history can identify a person even after obvious identifiers are removed. Treat the entire output folder as sensitive.

Back up the source folder yourself. The AI will not modify originals, but this workflow is not a backup system.

Personal use only. This workflow is written for individuals, friends, and family managing their own local documents.