AI workflow MCP-native Claude · ChatGPT · Gemini · MCP

The Agent Inbox

An agentic skill that turns one watched cloud folder into a tagged, searchable Notion archive of everything you save. Drop a screenshot, voice memo, PDF, article URL, or photo. Run process inbox. The skill reads each file in detail, proposes tags, files it. Over time it learns your taste and asks you fewer questions. Runs natively on Claude, ChatGPT, or Gemini.

Instant download for members · yours forever

What it does

  1. Watches Dropbox, Google Drive, or Mega.
  2. Reads images, screenshots, video, voice notes, PDFs, articles.
  3. Files into a searchable Notion database - along with tags, links & connections.
  4. Context awareness improves asset tagging over time.
  5. Can run autonomously on a regular routine (optional).
  6. Universal MCP support. Runs in Claude, ChatGPT, or Gemini.

How it works

  1. Pick a host. Claude Code, Claude.ai Projects, Claude Cowork / Desktop, ChatGPT Project or Custom GPT, or a Gemini Gem. Per-host install steps ship with the skill.
  2. Authorise the Notion MCP (or Notion connector) in your host. The skill writes to a database in your own Notion workspace.
  3. Pick one cloud source. Dropbox, Google Drive, or Mega. A synced local folder is fine where MCP is not available.
  4. Pick one folder inside that cloud as your inbox. Mostly empty is best. Open a chat. Say install agent-inbox. The skill walks you through the rest in about ten minutes.
  5. From then on you do two things only. Drop files in. Run process inbox. The skill reads, proposes, asks, files. Source files archive into processed/YYYY-MM/type/ after a successful Notion write.

Install

Our AI workflows are designed to be fully platform agnostic - meaning they are compatible with your AI platform of choice.

01 Claude Code MCP · recommended

One-line install. The skill registers as a project-scoped MCP server and is callable on any prompt that matches its description.

  1. Unzip the agent-inbox bundle anywhere on disk you can reference.
  2. From your project root, run the install command. Claude Code adds the server to .mcp.json and reloads automatically — no restart needed.
  3. Type /mcp in any Claude Code session to confirm the server is connected.
claude mcp add the-agent-inbox --command node --args ./the-agent-inbox/mcp/server.js
02 Claude Desktop MCP-native

Paste a server entry into your Claude Desktop config and the skill is available in every chat.

  1. Open claude_desktop_config.json. macOS: ~/Library/Application Support/Claude/claude_desktop_config.json. Windows: %APPDATA%\Claude\claude_desktop_config.json.
  2. Add the snippet below under mcpServers. If the file is empty, paste the full block.
  3. Quit and reopen Claude Desktop. The skill appears in the tools tray at the bottom of the input.
{
  "mcpServers": {
    "the-agent-inbox": {
      "command": "node",
      "args": ["/absolute/path/to/the-agent-inbox/mcp/server.js"]
    }
  }
}
03 Claude.ai Projects Paste-in

Claude.ai Projects don’t run MCP servers. Install as a Project: paste SKILL.md into Project Instructions, upload knowledge files into Project Knowledge.

  1. Create a new Project on claude.ai (or open an existing one).
  2. Open Project Instructions. Paste the full contents of SKILL.md from the bundle.
  3. Open Project Knowledge. Upload every file from /knowledge/ in the bundle (corpus files, templates, reference docs).
  4. Open a chat in the Project. Say "install agent-inbox" to start the walkthrough.
04 ChatGPT Project or Custom GPT

Paste-in path: build a Project or Custom GPT that wraps the skill. Use Custom GPT if you want it in your team’s GPT picker.

  1. In ChatGPT, go to Explore GPTs → Create.
  2. In the Configure tab, paste the full SKILL.md into Instructions.
  3. Upload every file from /knowledge/ in the bundle into the Knowledge field.
  4. Save. Open a chat with the new GPT and say "install agent-inbox" to start.
05 Gemini Gem

Create a Gemini Gem with the skill brain in its instructions field.

  1. Open Gemini → Gems → Create a new Gem.
  2. Paste the full SKILL.md into the Gem instructions field.
  3. Upload the knowledge files from /knowledge/ as Gem knowledge files.
  4. Save and open the Gem. Say "install agent-inbox" to start.

Stacks with

These AI workflows are designed to chain together. Run one's output through another and you get a real production pipeline.

  • AI Editor In Chief

    Already use AI Editor In Chief from MWM Issue 02? Install it alongside The Agent Inbox and conversational search summaries come back in your own writing voice.