mcp-storage-server
Storacha MCP storage server - self-sovereign data for your AI applications.
Documentation
Storacha MCP Storage Server
A Model Context Protocol (MCP) server implementation for Storacha hot storage, allowing AI applications to store and retrieve files through a standardized interface. It enables trustless, decentralized data exchange using IPFS and CIDs, ensuring data sovereignty, verifiability, and seamless integration with agent frameworks & AI systems.
Free Storage Options to Start ๐
Jump-start your decentralized storage journey today!
- GitHub Users: Sign up with your GitHub account and instantly receive 100MB of free storage โ no credit card required!
- Email Users: Register with your email and add a credit card to unlock 5GB of free storage space.
Use Cases
- Document Storage & Analysis: Securely upload and retrieve Blob documents.
- Long-term Structured Data Storage: Maintain structured data storage optimized for longevity and accessibility.
- Data Sharing Between Agents and Systems: Easily share data across multiple agents and diverse systems using CIDs (Content Identifiers), enabling decentralized, verifiable, and efficient data exchange.
- Application Integration: Seamlessly integrate Storacha storage retrieval into applications via the Model Context Protocol.
- AI Model Development: Support AI models by providing reliable versioning and access to external datasets stored in Storacha.
- LLM Integration: Enhance large language models (LLMs) by connecting directly with Storacha Storage for seamless data access.
- Web Application Backups: Reliably store backup copies of web applications for disaster recovery.
- Machine Learning Datasets: Efficiently manage and access large datasets used in machine learning workflows.
Quick Installation Guide
Get started with the Storacha MCP Storage Server in just a few simple steps.
1. Clone the Repository
git clone https://github.com/storacha/mcp-storage-server.git && cd mcp-storage-server2. Install Dependencies
pnpm install3. Generate Keys & Delegation
npm install -g @web3-storage/w3cliw3 login_Output:_
? How do you want to login?
Via Email
โฏ Via GitHub_Select Via GitHub and authenticate with your GitHub account._
w3 space create_Replace `` with a name for your new Space_. :warning: _Make sure you save the recovery key, so you can access your space from another device if needed._
w3 key create_Output:_
AgentId: did:key:z6MkhMZRW2aoJ6BQwkpMSJu68Jgqkz1FTpr1p69cpnN43YWG
PrivateKey: LMgCYLkvOc8Sm0mOL4cWFLxsWP0ZPEYrLxcQqsV93/s5RLje0BKx05muAse1Hkvh+sxUW38OcHtpiN1zxfpTJ4ht4jxV0=w3 delegation create \
--can 'store/add' \
--can 'filecoin/offer' \
--can 'upload/add' \
--can 'space/blob/add' \
--can 'space/index/add' --base64_Replace with the AgentId from the previous step. It grants the Agent the permission to store files into the recently created space_.
4. Configure the MCP Client
Next, configure your MCP client (such as Cursor) to use this server. Most MCP clients store the configuration as JSON in the following format:
{
"mcpServers": {
"storacha-storage-server": {
"command": "node",
"args": ["./dist/index.js"],
"env": {
// The server supports `stdio`, `sse`, and `rest` modes, the default is `stdio`.
"MCP_TRANSPORT_MODE": "stdio",
// The Storacha Agent private key that is authorized to store data into the Space.
"PRIVATE_KEY": "",
// The base64 encoded delegation that proves the Agent is allowed to store data. If not set, MUST be provided for each upload request.
"DELEGATION": "",
},
"shell": true,
"cwd": "./",
},
},
}_Replace `` with the PrivateKey you created in step 3. Then, replace the `` with the delegation you created in step 3._
### REST Mode and Cloud Hosting
The Storacha MCP Storage Server supports REST transport mode, which is compatible with MCP.so cloud hosting. To use REST mode:
{
"mcpServers": {
"storacha-storage-server-rest": {
"url": "http://localhost:3001/rest",
},
},
}For more information on deploying to MCP.so cloud, see the integrations.md guide.
_:warning: There are several ways to configure MCP clients, please read the integrations.md guide for more information._
Tools
The Storacha MCP Storage Server provides the following tools for AI systems to interact with a decentralized storage network.
Storage Operations
upload
Upload a file to the Storacha Network. The file must be provided as a base64 encoded string with a filename that includes the extension for MIME type detection.
interface UploadParams {
// Base64 encoded file content
file: string;
// Filename with extension for MIME type detection
name: string;
// Optional: Whether to publish to Filecoin (default: false)
publishToFilecoin?: boolean;
// Optional: Custom delegation proof
delegation?: string;
// Optional: Custom gateway URL
gatewayUrl?: string;
}retrieve
Retrieve a file from the Storacha Network. Supported filepath formats: `CID/filename`, `/ipfs/CID/filename`, or `ipfs://CID/filename`.
interface RetrieveParams {
// Path in format: CID/filename, /ipfs/CID/filename, or ipfs://CID/filename
filepath: string;
// Optional: Whether to use multiformat base64 encoding
useMultiformatBase64?: boolean;
}identity
Returns the `DIDKey` of the Storacha Agent loaded from the private key storage configuration.
interface IdentityParams {
// No parameters required
}See the integrations.md guide for detailed code examples and different integration patterns (SDK, Docker, etc).
License
MIT or Apache 2 License
Support
For support, please visit Storacha Support or open an issue in this repository.
Certifications
Frequently asked questions
What is mcp-storage-server?
mcp-storage-server is Storacha MCP storage server - self-sovereign data for your AI applications.
How do I install mcp-storage-server?
Open the GitHub repository and follow its README. Most MCP servers are added to your client's MCP config, then called by your agent.
Is mcp-storage-server open source?
Yes โ it is hosted on GitHub at https://github.com/storacha/mcp-storage-server and has 12 stars.
Related MCP tools
AI Agents & MCPs & AI Workflow Automation โข (~400 MCP servers for AI agents) โข AI Automation / AI Agent with MCPs โข AI Workflows & AI Agents โข MCPs for AI Ag...
๐ค A visualization mcp contains 25+ visual charts using @antvis. Using for chart generation and data analysis. TypeScript-based implementation.
Dive is an open-source MCP Host Desktop Application that seamlessly integrates with any LLMs supporting function calling capabilities. โจ
๐MCP server for accessing RedNote(XiaoHongShu, xhs). TypeScript-based implementation. Trusted by 800+ developers. Trusted by 800+ developers.
A Unified MCP Server Management App TypeScript-based implementation. Trusted by 800+ developers. Trusted by 800+ developers.
A Model Context Protocol (MCP) server implementation for remote memory bank management, inspired by Cline Memory Bank. TypeScript-based implementation.
Run your own MCP server? See who uses it and what to fix.
Measure it with TrackMCP