Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@shelby-mcp upload the meeting notes to Shelby as notes.txt"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
MCP server for Shelby Protocol — connect AI agents to decentralized storage.
Works with Claude Desktop, Cursor, Claude Code, Codex, and any MCP-compatible client.
AI agents can read, write, and manage files on Shelby's decentralized storage network (Aptos) through 5 tools:
Tool | Description |
| Upload text content to Shelby storage |
| Download a blob and return its content |
| List blobs under current account |
| Delete a blob |
| Show account info, balance, network |
Related MCP server: filesystem-mcp
Node.js v18+
Shelby CLI installed and configured (npm install -g @shelby-protocol/cli)
A funded Shelby account (shelby faucet)
git clone https://github.com/Jasekeee/shelby-mcp.git
cd shelby-mcp
npm install
npm run buildAdd to ~/Library/Application Support/Claude/claude_desktop_config.json:
{
"mcpServers": {
"shelby": {
"command": "node",
"args": ["/absolute/path/to/shelby-mcp/build/index.js"]
}
}
}Restart Claude Desktop. You can now say:
"Upload this report to Shelby as reports/q1-2026.txt"
"Download my config from Shelby"
"What's my Shelby account balance?"
Add to .cursor/mcp.json in your project:
{
"mcpServers": {
"shelby": {
"command": "node",
"args": ["/absolute/path/to/shelby-mcp/build/index.js"]
}
}
}claude mcp add shelby node /absolute/path/to/shelby-mcp/build/index.jsAI Agent (Claude/Cursor/Codex)
↕ MCP Protocol (stdio)
shelby-mcp server
↕ CLI wrapper
shelby CLI → Aptos blockchain → Shelby storage networkThe server wraps the shelby CLI rather than using the SDK directly. This ensures compatibility across CLI versions and avoids SDK breaking changes.
shelby-mcp reads from your existing ~/.shelby/config.yaml. No additional configuration needed — if shelby upload works in your terminal, it works through MCP.
PRs welcome. Please open an issue first for major changes.
MIT
Show current Shelby account info: name, address, balance, and active network context.
Inputs: no named parameters.
Delete a blob from Shelby storage.
Inputs: blob_name (required).
Download a blob from Shelby storage and return its content as text.
Inputs: blob_name (required).
List all blobs stored in Shelby under the current account, optionally filtered by prefix.
Inputs: prefix.
Upload a file to Shelby decentralized storage. Provide file content as text, a destination blob name, and expiration.
Inputs: content (required), blob_name (required), expiration.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 5 tools
Each tool has a distinct purpose: account info, delete, download, list, and upload. No two tools overlap in functionality, making selection unambiguous.
All tools use the 'shelby_' prefix followed by a verb, except 'shelby_account' which uses a noun but is still clear. The pattern is mostly consistent except for this minor deviation.
With five tools covering the core operations of a decentralized storage service (account, list, upload, download, delete), the count is well-scoped and manageable.
The tool set covers the main CRUD operations but lacks a metadata retrieval tool (e.g., blob size or creation time). This is a minor gap that agents can work around.
Persistent file storage for AI agents via MCP and curl. Upload, download, and version files.
Nifty's MCP server — exposes tasks, projects, messages, and files as tools for AI agents.
Persistent file storage for AI agents. Save files over MCP, get a link a human can open.
MCP server exposing the YourImageShare upload API (upload/list/delete) as tools for AI agents.
