🪐 MCP IPFS Server (storacha.network) 🛰️

A Node.js server implementing the Model Context Protocol (MCP) for interacting with the storacha.network platform via the w3 command-line interface (@web3-storage/w3cli).
This server empowers language models 🤖 and other MCP clients to manage storacha.network spaces, upload/download data, manage delegations, and perform various other tasks by seamlessly wrapping w3 commands.
✨ Features
- Wraps the
w3CLI for native integration with storacha.network. - Provides MCP tools covering a wide range of
w3functionality:- 🔑 Authentication & Agent:
w3_login,w3_reset,w3_account_ls(for checking authorization) - 📦 Space Management:
w3_space_ls,w3_space_use,w3_space_info,w3_space_add,w3_space_provision(Note:w3_space_createmust be run manually due to interactive prompts) - 💾 Data Management:
w3_up,w3_ls,w3_rm - 🔗 Sharing:
w3_open(generates w3s.link URL) - 🤝 Delegations & Proofs:
w3_delegation_create,w3_delegation_ls,w3_delegation_revoke,w3_proof_add,w3_proof_ls - 🔐 Keys & Tokens:
w3_key_create,w3_bridge_generate_tokens - ⚙️ Advanced Storage (
w3 can ...): Blob, CAR, Upload, Index, Access Claim, Filecoin Info management - 💳 Account & Billing:
w3_plan_get,w3_coupon_create,w3_usage_report
- 🔑 Authentication & Agent:
🛠️ Prerequisites
- Node.js: Version 22.0.0 or higher (
node -v). w3CLI: The server executesw3commands directly. Ensure@web3-storage/w3cliis installed globally and configured:npm install -g @web3-storage/w3cli w3 login <[email protected]> # Follow email verification steps- Environment Variable: The
w3_logintool requires theW3_LOGIN_EMAILenvironment variable to be set to the same email used forw3 login.
🏗️ Project Structure
The codebase is organized as follows:
src/
├── index.ts # Main server entry point, MCP setup, request routing
├── schemas.ts # Zod schemas defining input arguments for each tool
├── tool_handlers.ts # Implementation logic for each MCP tool
├── utils.ts # Helper functions (e.g., running w3 commands, parsing JSON)
└── utils/
└── logger.ts # Basic logger configuration
🚀 Usage with MCP Clients
This server can be used with any MCP-compatible client. You need to configure your client to connect to this server.
Example: NPX (Recommended for simple local use)
This assumes npm and the prerequisites are met.
{
"mcpServers": {
"ipfs": {
"command": "npx",
"args": ["-y", "mcp-ipfs"],
"env": {
"W3_LOGIN_EMAIL": "[email protected]"
}
}
}
}
Example: Docker
Build the image first (see Build section) or use the pre-built image alexbakers/mcp-ipfs.
{
"mcpServers": {
"mcp-ipfs": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-v",
"/path/to/your/project:/path/to/your/project",
"-e",
"W3_LOGIN_EMAIL",
"alexbakers/mcp-ipfs"
],
"env": {
"W3_LOGIN_EMAIL": "[email protected]"
}
}
}
}
📝 Note on Paths:
Several w3 commands require absolute filesystem paths (e.g., w3_up, w3_delegation_create --output, w3_proof_add, w3_can_blob_add, w3_can_store_add).
- NPX: Provide absolute paths from your host machine.
- Docker: Provide absolute paths inside the container. If interacting with files from your host (e.g., uploading), you must mount the relevant host directory into the container using the
-vflag (e.g.,-v /Users/me/project:/Users/me/project) and then use the container path (e.g.,/Users/me/project/my_file.txt) in the tool arguments.
📦 Build
Clone the repository and install dependencies:
git clone https://github.com/alexbakers/mcp-ipfs.git
cd mcp-ipfs
npm install
Build the TypeScript code:
npm run build
You can then run the server directly:
# Ensure W3_LOGIN_EMAIL is set in your environment
export W3_LOGIN_EMAIL="[email protected]"
node dist/index.js
Or publish it (if you have the rights):
npm publish
🐳 Docker Build
Build the Docker image:
# Build locally (replace with your username/repo and desired tag)
docker build -t alexbakers/mcp-ipfs .
📜 License
This MCP server is licensed under the MIT License. See the LICENSE file for details.
Recommend MCP Servers 💡
mcp-simple-pubmed
MCP server for searching and querying PubMed medical papers/research database
review-toolkit-mcp
A Model Context Protocol (MCP) service designed to assist AI agents in managing and tracking code review sessions, including file tracking, review status, agent review storage, and report generation.
hass-mcp
A Model Context Protocol (MCP) server for Home Assistant integration with Claude and other LLMs, enabling AI assistants to query, control, and summarize smart home entities.
@mastergo/magic-mcp
A standalone MCP service connecting MasterGo design tools with AI models to retrieve DSL data from design files.
TrueNasCoreMCP
Control TrueNAS Core system using natural language through MCP clients like Claude Desktop
reexpress_mcp_server
Reexpress MCP Server provides state-of-the-art statistical verification and an AI second opinion for complex LLM pipelines, particularly in software development and data science settings, by ensembling multiple LLMs and calculating predictive uncertainty.