File & Folder Organisation

Last verified: 14 February 2026 | Applies to: Pro, Max, Team, Enterprise (requires Cowork)

In 30 seconds

Every operator has a folder problem. Downloads full of cryptically named files, project folders that grew organically into chaos, shared drives where nobody can find anything. Claude in Cowork can read your file system, understand what files contain, and reorganise them — renaming, sorting, deduplicating, and restructuring based on actual content rather than just file names. It turns a weekend chore into a twenty-minute conversation.

Prerequisites

• Claude Desktop with Cowork enabled (Pro or above)
• A folder or directory you want to organise
• Back up the folder first — Cowork moves and renames real files

What it does

Claude in Cowork can read filenames, file metadata, and file contents within your working folder. This lets it do things a simple file manager cannot:

Task	What Claude does	Why it’s better than doing it manually
Semantic sorting	Reads file contents and sorts by topic, project, or type	Groups files by what they are about, not just what they are named
Bulk renaming	Renames files using a consistent convention	Processes hundreds of files with the same logic in seconds
Deduplication	Identifies duplicate or near-duplicate files	Catches renamed copies, versioned duplicates, and files saved in multiple locations
Restructuring	Creates a new folder hierarchy and moves files into it	Builds a logical structure based on content relationships
Archiving	Identifies old, unused, or irrelevant files	Separates active working files from archival material

How to set it up

Audit your current mess

Start by letting Claude understand what you are working with. Open a Cowork session, select the folder you want to organise, and ask for an audit.

Scan this folder and its subfolders. Give me a summary:
1. Total number of files, broken down by file type
2. Total size
3. Files with unclear or generic names (like "Document1.pdf" or "final_final_v3.docx")
4. Potential duplicates (same name, similar name, or same file size)
5. Files that haven't been modified in over 12 months
6. Any empty folders

Present this as a structured report. Don't move or change anything yet — just audit.

Claude scans the directory and produces an audit report. This gives you a clear picture before making any changes.

Semantic sorting

This is where Claude goes beyond what a normal file manager can do. Instead of sorting by name or date, Claude reads file contents and groups them by meaning.

Sort the files in this folder into logical groups based on their content. Read each file (or the first page for PDFs) and categorise them. Create subfolders for each category and move the files into them. Use clear, descriptive folder names.

Rules:
- Don't create more than 8 top-level categories (keep it simple)
- If a file could fit in multiple categories, put it in the most relevant one
- Create an "_unsorted" folder for anything you're not confident about
- Show me what you plan to do before executing — I want to approve the structure first

The “show me before executing” instruction is important. Claude proposes the new structure, you review it, and then Claude executes the moves. This prevents surprises.

Bulk renaming

Inconsistent file names make searching impossible. Claude can apply a naming convention across an entire folder.

Rename all files in this folder using this convention:
[YYYY-MM-DD]_[project-name]_[document-type]_[version].[ext]

Examples:
- 2026-02-14_acme-proposal_contract_v2.pdf
- 2026-01-30_internal_meeting-notes_v1.docx

Rules:
- Use the file's last-modified date if no date is apparent from the content
- Infer the project name from the file content or existing filename
- Use kebab-case (lowercase, hyphens between words)
- If you can't determine the project, use "general"
- Show me the proposed rename table before executing: current name → new name

Claude presents a rename table for your approval. Review it, make corrections, then let Claude execute.

Deduplication

Over time, files accumulate duplicates — copies saved to different folders, renamed versions, “backup” copies, and email attachments saved alongside originals.

Scan this folder tree for duplicate and near-duplicate files. Check for:
1. Exact duplicates (same content, different names or locations)
2. Version duplicates (same document, different versions — keep the newest)
3. Near-duplicates (files with very similar content that might be variations)

For each group of duplicates, show me:
- All file paths in the group
- Which one you recommend keeping (newest, most complete, or best-named)
- Why the others appear to be duplicates

Don't delete anything yet. Present your findings and let me decide what to remove.

Full restructure

For folders that need a complete overhaul — a shared drive that has been accumulating files for years, a project folder handed over from someone who left.

This folder contains everything from the Meridian project, which ran for 18 months. It's a mess — files at every level, no consistent naming, and probably duplicates. Restructure it into a clean archive:

Proposed structure:
- 01_contracts/ — all agreements, SOWs, amendments
- 02_financials/ — invoices, budgets, expense reports
- 03_deliverables/ — final versions of what we delivered
- 04_communications/ — important emails, meeting notes, status reports
- 05_working-files/ — drafts, working documents, internal notes
- 06_archive/ — anything superseded or no longer relevant

Move every file into the appropriate folder. Rename using the convention [date]_[description].[ext]. Create a README.md at the top level listing what's in each folder and the total file count.

How operators actually use it

What operators say

“Our shared drive had 4 years of unsorted files — about 3,000 documents across 200 folders. I pointed Claude at it and asked for an audit first. It found 340 duplicates and 600 files that hadn’t been touched in over two years. After a restructure that took about 45 minutes of back-and-forth, we had a clean, navigable folder structure. Would have taken someone a full week to do manually.”

“I use the bulk rename feature every month when our finance team exports reports. They come out of the system with names like ‘RPT_20260214_001.xlsx’. Claude renames them to something humans can read in about 30 seconds.”

Before this workflow: Shared drives are graveyards. Nobody can find anything. People create new versions instead of finding the existing one. The problem compounds every month.

After this workflow: Clean folder structures, consistent naming, no duplicates, and a clear separation between active and archived material. The audit step alone saves hours of searching.

What to watch out for

Caution

• Back up before you start. Cowork moves and renames real files on your file system. If Claude misfiles something or renames it incorrectly, you want a fallback. Copy the folder before starting, or use version control if available.
• Always preview before executing. Include “show me the plan before doing anything” in every prompt. Claude occasionally miscategorises files or proposes renames that lose important information from the original filename.
• Large directories take time. Reading the content of hundreds of files (especially PDFs) uses significant context and Cowork usage. For very large directories (1,000+ files), work in batches — one subfolder at a time.
• Claude cannot read all file types. Text files, PDFs, Word documents, spreadsheets, and markdown are well-supported. Proprietary formats, images, and binary files will be sorted by filename and metadata only, not content.
• Shared drives need coordination. If other people are actively working in the folder you are reorganising, files may move while they are editing them. Do major restructures during off-hours or on a copy of the folder, then swap it in.
• Cowork permissions are broad. When you select a working folder, Claude can read, write, move, and delete files in that folder and its subfolders. Be deliberate about which folder you select — do not point Cowork at your entire home directory.

Related

• Workplace Memory — use the Productivity plugin to maintain context about your file organisation conventions
• Cowork — understanding the sandboxed environment where file operations run
• Document Creation — once your files are organised, create new documents that follow the same structure

---

Something wrong or outdated? Let us know →

Get weekly workflows — subscribe to the newsletter.