agent-drop
Enables local file exchange between users and CLI agents via a web UI and MCP server, allowing agents to read/uploads and deliver artifacts without copy-paste.
README
Agent Drop
Local MCP file and secret handoff for CLI AI agents.
Agent Drop gives you a small local web UI plus a local MCP server so you and your CLI coding agents can exchange files, API keys, passwords, screenshots, and generated artifacts without pasting sensitive data into the chat transcript.
Use it with terminal agents such as Codex CLI, Gemini CLI, Claude Code, and other MCP-capable CLI tools.
You -> web UI -> user_drops/
Agent -> MCP -> agent_drops/
Your files stay on your machine.
Why
CLI agents are powerful, but file handoff is still awkward:
- You need to give the agent a screenshot, PDF, token, key, or log file.
- You need to pass an API key or password without leaving it in the session history.
- The agent needs to give you a generated report, image, patch, export, or artifact.
- Terminal sessions usually do not have a clean upload/download lane.
Agent Drop is the missing local dropbox between you and the agent.
Features
- Local web UI for uploading files and short notes to an agent.
- Secret-note flow for passwords, tokens, and API keys you do not want pasted into chat history.
- Download area for files delivered by agents.
- MCP server for CLI agents to read uploads and deliver artifacts.
- One shared local folder, configured with
AGENT_DROP_DIR. - One-time
.burnnotes for sensitive snippets. - Docker Compose setup for the web UI.
- Guided setup for local terminal environments.
- Optional reverse proxy support for advanced self-hosting.
Quick Start
Requirements:
- Node.js 20+
- npm
- Docker Engine with Docker Compose, or another Docker-compatible local runtime
Install:
git clone https://github.com/mikezio/agent-drop.git
cd agent-drop
npm run setup
docker compose up -d --build
Open:
http://localhost:8400/
The setup script prints MCP config snippets for your CLI agent.
Run Without Docker
npm --prefix secret-drop-ui install
npm --prefix mcp install
AGENT_DROP_DIR=./data/agent-drop npm run dev
Open http://localhost:3000/.
How It Works
Agent Drop uses one local directory:
AGENT_DROP_DIR/
user_drops/ # files and notes uploaded by you
agent_drops/ # files delivered by agents
The web UI and every MCP client must point to the same AGENT_DROP_DIR.
Good local defaults:
- Local terminal:
/home/you/AgentDrop - WSL/Linux VM:
/home/you/AgentDrop - Server:
/srv/agent-drop
Do not put AGENT_DROP_DIR in a public repo. It may contain secrets.
Configure Your CLI Agent
Agent Drop has one MCP server in mcp/.
MCP command:
/absolute/path/to/agent-drop/mcp/run.sh
Codex CLI TOML:
[mcp_servers.agent-drop]
command = "/absolute/path/to/agent-drop/mcp/run.sh"
env = { AGENT_DROP_DIR = "/absolute/path/to/AgentDrop" }
Generic MCP JSON for Gemini CLI, Claude Code, and similar clients:
{
"mcpServers": {
"agent-drop": {
"command": "/absolute/path/to/agent-drop/mcp/run.sh",
"env": {
"AGENT_DROP_DIR": "/absolute/path/to/AgentDrop"
}
}
}
}
More examples: docs/agent-clients.md
MCP Tools
get_unread_uploads: returns unread user uploads and marks them as read.list_all_uploads: lists all user uploads.deliver_to_user_device: copies generated files intoagent_drops.generate_and_deliver_file: creates a short text file inagent_drops.delete_specific_drop: deletes one file from either drop folder.clear_drops: clears user drops, agent drops, or both.
Local First
Agent Drop is intentionally local-first:
- No account.
- No cloud service.
- No database.
- No hosted file storage.
- No domain required.
If you want LAN access, a domain, or /drop behind nginx/Caddy/Traefik, see docs/deployment.md.
Security
Treat your drop directory as sensitive storage.
- Keep the UI bound to
127.0.0.1unless you intentionally want network access. - Add authentication before exposing the UI to a LAN or public domain.
- Secret notes avoid chat-history exposure; they are still local files until read or cleared.
- Never commit
.env,data/, or dropped files. - Clear old drops when a project is done.
See SECURITY.md.
Development
npm --prefix secret-drop-ui install
npm --prefix mcp install
AGENT_DROP_DIR=./data/agent-drop npm run dev
Build:
npm --prefix secret-drop-ui run build
Run the MCP server:
AGENT_DROP_DIR=./data/agent-drop npm --prefix mcp run start
Project Layout
secret-drop-ui/ Next.js web UI
mcp/ stdio MCP server
scripts/ guided setup and optional utilities
docs/ client and deployment notes
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.