What can you do with this server?

The shoonya-mcp server is a secure Node.js-based MCP interface for algorithmic trading via the Shoonya Broker API, designed for seamless integration with AI assistants like Claude, Cursor, or Gemini. Secure Credential Management: Encrypts broker credentials using AES-256-GCM with PBKDF2 (310,000 iterations), storing them locally in an encrypted vault (~/.shoonya-mcp/vault.enc). The AI never directly sees sensitive credentials. Includes CLI commands for one-time setup, editing (e.g., for 90-day password resets), and securely deleting the vault. Core Trading Capabilities: * Login — Authenticate using only a master password; credentials are decrypted locally and never exposed to the AI. * Place Standard Orders — Place market or limit orders via a simple command string (e.g., NIFTY 24500 CE 30 L 110). * Place Manual Stop Loss Orders — Place SL orders using a command string (e.g., NIFTY 24500 CE 30 SL 100). * Place Auto Stop Loss Orders — Automatically fetch the average buy price of a filled order and place an SL-LMT order based on index-specific offsets. * Modify Open Orders — Update price, quantity, or price type of an existing open order. * Modify Stop Loss Orders — Adjust price, trigger price, quantity, or price type of an existing SL order. * Cancel Orders — Cancel any open order, including stop loss orders. * Exit Positions — Exit an open position by placing an opposite market order. * View Order Book — Fetch and review your full order book. * Check Margin — Retrieve available trading margin (cash + payin + payout − margin used). * Check Order Status — Look up the current status of a specific order by order ID.

How do I use shoonya-mcp?

1. Click on "Deploy Server". 2. Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state. 3. In the chat, type @ followed by the MCP server name and your instructions, e.g., "@shoonya-mcp Buy NIFTY 24500 CE 30" 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.

shoonya-mcp

by benatcmathew

Overview Schema Source Related Servers Discussions Admin

TypeScript

Local

Shoonya MCP Server (Node.js)

A lightweight, bank-level secure Model Context Protocol (MCP) server for algorithmic trading via the Shoonya Broker API.

This package is designed for effortless and secure distribution. It runs natively over stdio, meaning it can be plugged directly into AI assistants like Claude Desktop, Cursor, or Gemini without requiring a Python environment.

🔐 Security Architecture

Your trading credentials are protected using military-grade encryption. The AI assistant never sees, stores, or transmits your passwords.

How It Works

┌─────────────────────────────────────────────────────────┐
│                   YOUR LOCAL MACHINE                    │
│                                                         │
│  ┌──────────────┐    Master     ┌──────────────────┐    │
│  │  AI (Claude,  │   Password   │  Shoonya MCP     │    │
│  │  Cursor, etc) │ ──────────►  │  Server (Node.js)│    │
│  │               │              │                  │    │
│  │  Never sees   │              │  Decrypts vault  │    │
│  │  credentials  │              │  locally only    │    │
│  └──────────────┘              └────────┬─────────┘    │
│                                         │               │
│                                         ▼               │
│                              ┌──────────────────┐       │
│                              │  ~/.shoonya-mcp/  │       │
│                              │  vault.enc        │       │
│                              │                   │       │
│                              │  AES-256-GCM      │       │
│                              │  encrypted file   │       │
│                              └──────────────────┘       │
│                                         │               │
│                                         ▼               │
│                              ┌──────────────────┐       │
│                              │  Shoonya Broker   │       │
│                              │  API (HTTPS)      │       │
│                              └──────────────────┘       │
└─────────────────────────────────────────────────────────┘

Encryption Details

Feature	Specification
Encryption Algorithm	AES-256-GCM (used by banks, governments, military)
Key Derivation	PBKDF2 with SHA-256, 310,000 iterations (OWASP 2023)
Salt	32 bytes, cryptographically random, unique per vault
IV (Nonce)	16 bytes, cryptographically random, unique per save
Authentication	GCM auth tag prevents tampering
File Permissions	`0600` (owner read/write only, Unix systems)
Vault Location	`~/.shoonya-mcp/vault.enc` (never in your project)
AI Exposure	❌ Credentials never sent to or seen by the AI

What the AI Can See vs Cannot See

Data	AI Can See?
Master Password	✅ Only to decrypt vault (not stored)
User ID	❌ Never
Broker Password	❌ Never
TOTP Secret Key	❌ Never
Client ID (formerly Vendor Code)	❌ Never
Secret Code (formerly API Key)	❌ Never
IMEI	❌ Never
Order commands	✅ (e.g., "NIFTY 24500 CE 30")
Order results	✅ (order ID, margin, status)

Related MCP server: Phantom MCP

Quick Start

Step 1: Setup Credentials (One-Time)

Run the interactive setup wizard in your terminal:

npx github:benatcmathew/Shoonya_Trading_MCP-Server --setup

This will:

Ask for your Shoonya broker credentials
Ask you to set a master password
Encrypt everything with AES-256-GCM and save to ~/.shoonya-mcp/vault.enc

Editing Credentials (90-Day Password Reset)

Shoonya requires users to reset their broker password every 90 days. To update your vault with your new password (or edit any other API details), simply run the edit command:

npx github:benatcmathew/Shoonya_Trading_MCP-Server --edit

When prompted, type yes to overwrite your existing vault, and enter your new details.

Step 2: Connect to Claude Desktop

Add this to your claude_desktop_config.json:

{
  "mcpServers": {
    "shoonya-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "github:benatcmathew/Shoonya_Trading_MCP-Server"
      ]
    }
  }
}

Step 3: Start Trading!

In Claude Desktop, simply say:

"Login to Shoonya" → The AI will ask for your master password only. "Buy NIFTY 24500 CE 30" → Places a market order. "Place auto SL for order 12345 on NIFTY" → Auto stop-loss.

CLI Commands

Command	Description
`npx shoonya-mcp-server --setup`	First-time credential setup (interactive wizard)
`npx shoonya-mcp-server --edit`	Edit/update stored credentials
`npx shoonya-mcp-server --delete`	Securely wipe and delete the vault
`npx shoonya-mcp-server --status`	Check if a vault exists
`npx shoonya-mcp-server`	Start the MCP server (default, used by Claude)

Available MCP Tools

Tool	Parameters	Description
`login`	`master_password`	Decrypt vault & login to Shoonya
`place_order`	`buy_or_sell`, `command`	Smart order (`buy_or_sell` must be `"B"` or `"S"`). Limit Orders ONLY for Options (Market orders are blocked via API). Use `"NIFTY 24500 CE 30 L 110"`
`place_manual_sl`	`command`	Manual SL (e.g., `"NIFTY 24500 CE 30 SL 100"`)
`place_auto_sl`	`order_id`, `index`	Auto SL from filled order
`modify_order`	`order_id`, `exchange`, `tradingsymbol`, ...	Modify open order
`modify_sl`	`order_id`, `exchange`, `tradingsymbol`, ...	Modify SL order
`cancel_order`	`order_id`	Cancel any order
`exit_order`	`order_id`, `product_type`	Exit position
`get_order_book`	—	Fetch order book
`check_margin`	—	Fetch available margin
`check_order_status`	`order_id`	Check specific order status

Manual Build from Source

git clone https://github.com/benatcmathew/Shoonya_Trading_MCP-Server.git
cd Shoonya_Trading_MCP-Server
npm install
npm run build
npm start

License

ISC

Available Tools

11 tools

cancel_orderB

Cancel any open order (including SL).

Inputs: order_id (required).

check_marginA

Fetch the available margin (cash + payin + payout - marginused).

Inputs: no named parameters.

check_order_statusC

Fetch the status of a specific order from the order book.

Inputs: order_id (required).

exit_orderC

Exit an open position by placing an opposite market order.

Inputs: order_id (required), product_type (required).

get_order_bookB

Fetch the user's order book.

Inputs: no named parameters.

modify_orderC

Modify an existing open order.

Inputs: exchange (required), order_id (required), tradingsymbol (required), price, +2 more.

modify_slC

Modify an existing Stop Loss order.

Inputs: exchange (required), order_id (required), tradingsymbol (required), price, +3 more.

place_auto_slA

Automatically fetch avg buy price of a filled order and place an auto SL-LMT order based on index specific offset.

Inputs: index (required), order_id (required).

place_manual_slB

Place a manual Stop Loss order using a string command like 'NIFTY 24500 CE 30 SL 100'

Inputs: command (required).

place_orderB

Place a standard order (Market/Limit) using a command string like 'NIFTY 24500 CE 30 L 110'

Inputs: command (required), buy_or_sell (required).

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

Deploy Server

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

TDQS

B3.3/5.0

Scored 2026-07-16 20:07 across 11 tools

Disambiguation5/5

Each tool has a distinct purpose: login, various order placement (standard, auto SL, manual SL), modifications, cancellation, exit, margin/status checks. No overlapping functions.

Naming Consistency4/5

Most tools follow a verb_noun pattern (e.g., cancel_order, place_order, check_margin). 'login' is a single verb but is a clear exception.

Tool Count5/5

11 tools cover the core trading lifecycle from authentication to order management and status checks, without being excessive.

Completeness4/5

Covers login, order placement (3 types), modification (2 types), cancellation, exit, margin, and order book. Missing a position book view and trade history, but core operations are well-covered.

Maintenance

ActivityInactive

ResponsivenessNo recent community issues

Resources

Related MCP Connectors

BotSpot: AI Trading, Backtests and BotsB
io.github.grzesir
AI agents backtest, paper trade and live trade via your broker. Free to connect; OAuth.
Finance Autonomous Agents AI & Machine Learning
1152
https://mcp.botspot.trade/mcp
Tradetron
tech.tradetron.mcp
Build, backtest and diagnose algo trading strategies in your own Tradetron account in plain English.
Finance App Automation
1
https://mcp.tradetron.tech/mcp
Trade It
app.tradeit
Connect brokerage accounts to AI assistants: holdings, orders, asset lookup, and approved trading.
Finance Cryptocurrency E-commerce & Retail
162
https://mcp.tradeit.app/mcp
Ice Juice Trading
com.icejuicetrading
Automate trading on your own Alpaca account - build, backtest and run strategies via your AI.
Finance Autonomous Agents
1
https://mcp.icejuicetrading.com/mcp

Related MCP Servers

AWS FinOps MCP Server
Cloud Platforms Finance
ravikiranvm
A
license
Not graded
quality
D
maintenance
Enables natural language analysis of AWS costs, automated FinOps waste audits, and budget monitoring across multiple profiles and regions while keeping credentials secure locally.
Updated a year ago (2025-06-11 13:53 UTC)
25 PyPI
184
MIT
Phantom MCP
Penetration Testing Security
mokhalifa83
F
license
Not graded
quality
D
maintenance
Enables authorized compliance verification and security auditing through natural language, bridging AI assistants with industry-standard security tools for enterprise audits.
Updated 10 months ago (2026-01-16 02:02 UTC)
24
-
ssh-licco
Shell Access Command Line
Echoqili
A
license
A
quality
B
maintenance
Enables natural language control of SSH servers, allowing users to execute commands, manage files, view logs, deploy applications, and more through AI-driven conversations.
Updated 15 days ago (2026-10-26 16:05 UTC)
9
73 PyPI
4
MIT
AWS FinOps MCP Server
Cloud Platforms Finance
Yukis6040
F
license
Not graded
quality
A
maintenance
Enables AI assistants to analyze AWS cloud costs, audit for waste, monitor budgets, and generate FinOps reports using natural language across multiple AWS profiles and regions, while keeping credentials on the local machine.
Updated 25 days ago (2026-10-16 09:53 UTC)
-