Building an AI-Powered Post-Production Pipeline with Handrive + Claude

This guide shows you how to build a complete AI-powered file pipeline for post-production. We'll use Claude Code (or Claude Desktop) as the AI agent and Handrive's MCP server for file operations.

By the end, you'll have an AI assistant that can:

• Monitor folders for new camera footage
• Automatically create shares with proper naming
• Add team members with appropriate roles
• Organize files using your naming conventions
• Notify team members when files are ready

Prerequisites

• Handrive installed and authenticated
• Claude Desktop or Claude Code with MCP enabled
• Basic familiarity with command line

Step 1: Configure Claude Desktop for Handrive MCP

First, add Handrive to your Claude Desktop configuration:

macOS

# Edit the config file
nano ~/Library/Application\ Support/Claude/claude_desktop_config.json

Windows

# Edit the config file
notepad %APPDATA%\Claude\claude_desktop_config.json

Configuration Content

{
  "mcpServers": {
    "handrive": {
      "command": "/path/to/handrive",
      "args": ["mcp"]
    }
  }
}

Replace /path/to/handrive with:

• macOS (app): /Applications/Handrive.app/Contents/MacOS/handrive
• macOS (binary): /usr/local/bin/handrive
• Windows: C:\\Program Files\\Handrive\\handrive.exe

Restart Claude Desktop after saving.

Step 2: Verify the Connection

Open Claude Desktop and ask:

Claude should use the list_shares MCP tool and return your shares. If this works, the connection is configured correctly.

Step 3: Build the Automated Ingest Pipeline

Now we'll create an AI-powered ingest workflow. Here's the complete prompt you can give to Claude:

I want you to help me set up an automated dailies ingest workflow.

When I tell you that new footage has arrived in /Volumes/DIT_Drive:

1. List the folders in /Volumes/DIT_Drive to find the new camera card
2. Parse the folder name to extract:
   - Shoot date (format: YYYY-MM-DD)
   - Camera (A-Cam, B-Cam, etc.)
3. Create a Handrive share named "Dailies - [Date] - [Camera]"
4. Add these team members:
   - editor@studio.com with Editor role
   - director@studio.com with Viewer role
   - producer@studio.com with Viewer role
5. Report back what was created

Use Handrive MCP tools for all file and share operations.

When you trigger this (by saying "New footage arrived in the DIT drive"), Claude will execute the entire workflow using MCP tools.

Step 4: Add File Organization

Extend the workflow to organize files:

After creating the share, organize the files:

1. List all files in the camera card folder
2. For each clip:
   - Parse the filename to extract scene and take numbers
   - Create a folder structure: Scene_XX/Take_XX/
   - Move the clip to the appropriate folder
3. Create a manifest.txt listing all clips with their locations
4. Report the final folder structure

Example: Complete Multi-Department Pipeline

Here's a more complex example that handles the entire post-production flow:

Help me manage a complete post-production pipeline:

## Dailies Ingest
When new camera cards arrive:
- Create dated share, add editor + director
- Organize by scene/take
- Generate clip manifest

## Edit Handoff
When I say "Picture lock for Episode X":
- Find the Editor's export in the Exports share
- Create "VFX Handoff - Episode X" share
- Add VFX supervisor with Editor role
- List VFX shot requirements (I'll provide these)

## VFX Returns
When VFX delivers:
- Verify frame counts match expected
- Create "Color - Episode X" share
- Add colorist with Editor role
- Remove VFX team access
- Notify colorist

## Color Delivery
When color is approved:
- Create "Final Delivery - Episode X" share
- Add delivery coordinator with Viewer role
- Generate delivery manifest with specs

Use Handrive MCP tools for all operations. Track progress
and report status after each step.

MCP Tools You'll Use

Here are the key Handrive MCP tools for pipeline automation:

Why This Beats Watch Folders

Traditional watch folders do one thing: detect new files and trigger a fixed action. They can't:

• Parse file names to understand context
• Make decisions about routing
• Apply different rules for different content types
• Handle multi-step workflows
• Adapt when something unexpected happens

AI agents with Handrive MCP can do all of these. The AI understands your production workflow and makes intelligent decisions about how to handle each file.

Extending the Pipeline

Once you have the basics working, extend with:

• Slack/email notifications: Have Claude send messages when files are ready (using external APIs)
• Automated QC: Check frame counts, codec, resolution before moving files to next stage
• Version tracking: Maintain a log of all versions and who accessed what
• Storage management: Archive completed projects, manage disk space

Tips for Production Use

1. Test on non-critical projects first. Build confidence in the automation before deploying on active productions.
2. Keep Claude's prompts in a document. Version control your automation prompts like code.
3. Add confirmation steps. For destructive operations (removing members, deleting files), have Claude ask for confirmation.
4. Monitor the logs. Review what the AI is doing until you're confident in the automation.