About million
An MCP server for operating million, a self-hosted Solana whale tracker and copy-trading deck, through its local API. It lets users inspect positions and trading decisions, manage rules and wallets, and check tokens; it cannot switch paper trading to live.
Key features
- 01Inspect positions and trading decisions
- 02Read and update trading rules
- 03Track wallet flows and holdings
- 04Check token safety
- 05Analyze and manage wallet roster
Use cases
- 01Understanding why a token was or was not bought
- 02Reviewing positions and exit reasons
- 03Finding and following wallets from token buyers
- 04Checking token risks before trading
MCP support
Provides MCP clients with tools to read and operate the running million deck through its local API. The deck must be running on the same machine; the server does not pass through keys or enable live trading.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
Tools
14Saved tool list · updated 2w ago
This tool takes no parameters.
Details
From the project READMEmillion-mcp
MCP server for million, the open-source, self-hosted Solana whale tracker and copy-trading deck. It lets Claude Desktop, Cursor, Claude Code, VS Code or any MCP client read and operate your deck: status, positions, why a token was or wasn't bought, trading rules, roster order flow, token safety checks and finding wallets worth following.
It's a thin layer over the deck's local API, so the deck has to be running on your machine. No keys pass through it, and no tool can switch paper trading to live.
1. Run the deck
npx create-million
That clones, installs and starts million at http://localhost:5173 (you need Node 24 or nvm, and a free Helius API key).
2. Add the server to your client
Claude Code
claude mcp add million -- npx -y million-mcp
Claude Desktop (claude_desktop_config.json), Cursor (.cursor/mcp.json) and most other clients:
{
"mcpServers": {
"million": {
"command": "npx",
"args": ["-y", "million-mcp"]
}
}
}
VS Code (.vscode/mcp.json):
{
"servers": {
"million": { "type": "stdio", "command": "npx", "args": ["-y", "million-mcp"] }
}
}
If the deck's API isn't on http://localhost:3001/api, set MILLION_API in the server's env.
Tools
| Tool | What it does |
|---|---|
get_status | running or not, paper or live, feed health, followed wallets, book summary |
get_book | open and closed positions with PnL and exit reasons |
get_decisions | the decision log: signals, trades, shadow trades and skips, each with its reason |
why_not_bought | everything to explain one token: its decisions, check report, positions, rules |
get_rules | trading rules in force |
update_rules | change some rule fields, returns before and after (can't enable live trading) |
get_flows | tokens the followed wallets are accumulating or distributing now |
get_roster_holdings | what the followed wallets still hold, and where they took profit |
check_token | safety checks: authorities, liquidity, holders, deployer, sell simulation, RugCheck |
find_token_buyers | a token's biggest buyers with flags (bots, clusters), to find whales to follow |
import_wallets | add wallets to the roster |
analyze_wallet | a wallet's PnL in SOL, win rate, hold times and flags |
get_wallet | a roster wallet's stored analysis and activity |
set_wallet_subscription | follow or unfollow a wallet |
find_token_buyers and analyze_wallet spend your Helius credits.
Ask things like
- "How's my deck doing?"
- "Why didn't it buy CATE?"
- "Find whales from this token and add the three most consistent ones."
- "Make it only trade when two different owners agree, 0.2 SOL per trade."
Experimental software, provided as is, not financial advice. MIT licensed.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.