Skip to main content
Glama
bahi-mcp

bahi-mcp

bahi-mcp

MCP server that gives AI assistants (Claude, Cursor, VS Code and others) access to Indian mutual fund data, NSE/BSE market data and financial calculators.

Bahi (बही) is the traditional Indian ledger. This server keeps the numbers straight so your assistant doesn't guess them.

Disclaimer: bahi-mcp provides information and calculations from public data sources (AMFI via the mftool library, and Yahoo Finance for stocks) and from assumptions you supply. It is not investment or tax advice and the author is not a SEBI-registered adviser. Data may be delayed or inaccurate, and tax rules change. Verify before making decisions.

What you can ask

  • "Compare these three flexi-cap funds on 3-year return, volatility and max drawdown."

  • "How often has this fund beaten 12% on rolling 3-year returns?"

  • "If I had invested ₹10,000 a month in scheme 119551 since January 2020, what would it be worth today, and what is my XIRR?"

  • "What's the Nifty at, and how has RELIANCE done over the last year?"

  • "How much tax would I pay if I sell this equity fund after 14 months with a ₹2 lakh gain?"

  • "How much corpus do I need to retire at 55 if I spend ₹60,000 a month today?"

Related MCP server: Fund Dashboard MCP Server

Tools

Mutual fund data (AMFI)

Tool

What it does

search_schemes

Search schemes by name, returns AMFI scheme codes

get_scheme_quote

Latest NAV and quote for a scheme code

get_scheme_details

Fund house, type, category, start date

get_nav_history

NAV history with optional date range (capped to keep responses small)

get_portfolio_value

Current value of a list of holdings (scheme code + units)

clear_cache

Drop cached AMFI data so the next call fetches fresh NAVs

Mutual fund analytics

Tool

What it does

calculate_sip_returns

Monthly SIP simulation on a fund's real NAV history: invested, units, value, absolute return, XIRR

compare_funds

Side-by-side 1/3/5-year returns, volatility, max drawdown and Sharpe ratio for 1 to 5 funds

get_rolling_returns

Distribution of rolling returns: min, max, mean, median, how often positive or above a threshold

Stocks and market data (NSE/BSE, via Yahoo Finance; unofficial and best-effort)

Tool

What it does

get_stock_quote

Latest price, day and 52-week range for an NSE/BSE stock

get_index_quote

Nifty 50, Sensex, Bank Nifty, Nifty IT, India VIX

get_stock_history

OHLCV history for a stock or index

Calculators

Tool

What it does

calculate_xirr

XIRR for any dated cash flows

calculate_cagr

CAGR between two values over a date range

calculate_emi

Loan EMI, total payment, total interest

calculate_fd / calculate_rd

Fixed and recurring deposit maturity

calculate_lumpsum_value

Future value of a one-time investment

project_sip

SIP projection at an assumed return, with optional yearly step-up

calculate_swp

Does a corpus last under a monthly withdrawal?

plan_retirement

Corpus needed and the monthly SIP to build it

estimate_mf_tax

Capital gains tax estimate on a fund redemption; use purchase/sale dates for exact 12-month classification

Dates can be DD-MM-YYYY or YYYY-MM-DD.

Install

Requires Python 3.10+.

pip install bahi-mcp                # mutual funds + calculators
pip install "bahi-mcp[stocks]"      # adds NSE/BSE stock and index data

Or run without installing, using uv:

uvx bahi-mcp
uvx --with yfinance bahi-mcp        # with stock data

Configure your MCP client

Claude Desktop

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "bahi-mcp": {
      "command": "uvx",
      "args": ["--with", "yfinance", "bahi-mcp"]
    }
  }
}

Drop "--with", "yfinance" if you don't need stock data. Restart Claude Desktop.

Cursor and other clients

{
  "mcpServers": {
    "bahi-mcp": {
      "command": "uvx",
      "args": ["--with", "yfinance", "bahi-mcp"]
    }
  }
}

For Cursor, save this in ~/.cursor/mcp.json. Remove "--with", "yfinance" if you don't need stock data.

Notes on data and accuracy

  • Mutual fund data comes from AMFI through mftool (3.4 or newer). NAVs are published after market close, usually by about 9 PM IST. mftool caches NAVs for 24 hours, so use clear_cache after the evening update to see the latest.

  • Stock data comes from Yahoo Finance through yfinance. It is unofficial, can be delayed, rate-limited or change without notice. Don't use it for trading decisions.

  • Tax estimates follow the configured FY 2026-27 assumptions for equity funds (STCG 20%, LTCG 12.5% above the remaining ₹1.25 lakh annual exemption); provide purchase and sale dates for exact calendar 12-month classification, or holding days for an approximate 365-day cutoff. The exemption is shared across eligible gains in a financial year. Debt funds are modelled at slab rate only for units bought on/after 1 April 2023. Surcharge, loss set-off and older indexation rules are not modelled; verify current law and consult a tax professional.

  • Projections (project_sip, calculate_swp, plan_retirement) assume constant returns. Real markets don't behave that way.

Development

git clone https://github.com/NayakwadiS/bahi-mcp
cd bahi-mcp
python -m venv .venv
source .venv/bin/activate        # Windows: .venv\Scripts\activate
pip install -e ".[dev,stocks]"
pytest
python -m bahi_mcp               # runs the server over stdio

Layout in src/bahi_mcp/:

  • analytics.py, fund_analytics.py, calculators.py: pure Python, no network, fully tested offline

  • data.py: AMFI access through mftool

  • market.py: stock and index access through yfinance

  • server.py: the MCP tool definitions

Roadmap

  • Portfolio overlap and fund-to-fund correlation

  • Document search over factsheets and SEBI/RBI circulars

  • Hosted version with caching

Contributing

Issues and pull requests are welcome.

License

MIT. See LICENSE.

Available Tools

22 tools
calculate_cagrC

CAGR between two values over a date range.

Inputs: end_date (required), start_date (required), final_value (required), initial_value (required). Returns: object

calculate_emiB

Loan EMI, total payment and total interest (reducing balance).

Inputs: years (required), principal (required), annual_rate_pct (required). Returns: object

calculate_fdA

Fixed deposit maturity amount and effective yield (quarterly compounding by default).

Inputs: years (required), principal (required), annual_rate_pct (required), compounding_per_year. Returns: object

calculate_lumpsum_valueB

Future value of a one-time investment at a constant assumed annual return.

Inputs: years (required), principal (required), expected_return_pct (required). Returns: object

calculate_rdC

Recurring deposit maturity amount (quarterly compounding).

Inputs: months (required), annual_rate_pct (required), monthly_deposit (required). Returns: object

calculate_sip_returnsA

Simulate a monthly SIP in a scheme using its real NAV history.

Returns total invested, units, current value, absolute return and XIRR. end_date defaults to today. Dates as DD-MM-YYYY or YYYY-MM-DD.

Inputs: start_date (required), scheme_code (required), monthly_amount (required), end_date. Returns: object

calculate_swpC

Systematic withdrawal plan: does the corpus last, and how much remains?

Inputs: years (required), corpus (required), monthly_withdrawal (required), expected_return_pct (required). Returns: object

calculate_xirrA

XIRR for dated cash flows.

Each item: {"date": "YYYY-MM-DD", "amount": -10000}. Investments are negative, redemptions / current value positive.

Inputs: cashflows (required). Returns: object

clear_cacheA

Clear cached AMFI data so the next call fetches fresh NAVs.

NAVs are cached for 24 hours. AMFI usually publishes the day's NAVs after about 9 PM IST, so clear the cache after that to see them.

Inputs: no named parameters. Returns: object

compare_fundsA

Compare 1 to 5 mutual funds side by side.

For each fund: latest NAV, 1/3/5-year annualised returns (null if the fund is too new), and risk over risk_years: volatility, max drawdown and Sharpe ratio.

Inputs: scheme_codes (required), risk_years, risk_free_rate_pct. Returns: object

estimate_mf_taxA

Estimate capital gains tax on a mutual fund redemption (FY 2026-27 rules).

Supply holding_days, or both purchase_date and sale_date for exact calendar holding-period classification. Equity STCG is 20%; LTCG is 12.5% above the remaining annual exemption. 'debt'/'other': taxed at slab_rate_pct. Excludes surcharge and loss set-off. Confirm with a tax professional.

Inputs: sale_amount (required), purchase_amount (required), fund_type, sale_date, +4 more. Returns: object

get_index_quoteA

Latest level of an Indian index: nifty50, sensex, banknifty, niftyit or indiavix.

Inputs: index. Returns: object

get_nav_historyA

NAV history for a scheme. Dates as DD-MM-YYYY or YYYY-MM-DD.

Returns at most max_points of the most recent points in the range, to keep responses small.

Inputs: scheme_code (required), end_date, max_points, start_date. Returns: object

get_portfolio_valueB

Current value of a mutual fund portfolio at latest NAVs.

Each holding: {"scheme_code": "119551", "units": 125.5}.

Inputs: holdings (required). Returns: object

get_rolling_returnsA

Rolling-return distribution for a fund.

Shows min, max, mean and median annualised return across all window_years windows in its history, plus how often it was positive and above threshold_pct. Better than a single start-to-end return.

Inputs: scheme_code (required), step_days, window_years, threshold_pct. Returns: object

get_scheme_detailsC

Fund house, scheme type, category and start date for a scheme code.

Inputs: scheme_code (required). Returns: object

get_scheme_quoteA

Latest NAV and basic quote for a scheme, by AMFI scheme code.

Inputs: scheme_code (required). Returns: object

get_stock_historyA

Price history (OHLCV) for an NSE/BSE stock or index ticker like '^NSEI'.

period: 1d, 5d, 1mo, 3mo, 6mo, 1y, 2y, 5y, 10y, ytd, max. interval: 1d, 5d, 1wk, 1mo. Returns at most max_points recent rows.

Inputs: symbol (required), period, exchange, interval, +1 more. Returns: object

get_stock_quoteA

Latest price for an NSE/BSE stock, e.g. symbol 'RELIANCE' on exchange 'NSE'.

Source is Yahoo Finance (unofficial, may be delayed).

Inputs: symbol (required), exchange. Returns: object

plan_retirementC

Corpus needed at retirement and the monthly SIP required to build it.

Inputs: current_age (required), retirement_age (required), life_expectancy (required), current_monthly_expense (required), +4 mor… Returns: object

project_sipA

Project a SIP at an assumed constant return, with an optional yearly step-up.

This is a what-if projection. For a SIP on a real fund's history, use calculate_sip_returns.

Inputs: years (required), monthly_amount (required), expected_return_pct (required), annual_step_up_pct. Returns: object

search_schemesA

Search Indian mutual fund schemes by name (e.g. 'sbi midcap direct growth').

Returns scheme codes (AMFI) and names. Use a code with the other tools.

Inputs: query (required), limit. Returns: result

Tool Schema Changelog

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

Loading tool schema changelog…

    TDQS

    A3.5/5.0

    Scored across 22 tools

    Disambiguation5/5

    Each tool targets a distinct action/resource: scheme search, scheme data, NAV history, portfolio valuation, mutual-fund returns, tax, calculators, and stock/index quotes. The closest overlap is project_sip versus calculate_sip_returns, but the descriptions explicitly separate what-if SIP projection from real-NAV SIP simulation. No tool pair appears interchangeable.

    Naming Consistency5/5

    All tool names use consistent snake_case verb_noun patterns such as get_scheme_quote, calculate_xirr, compare_funds, and get_stock_history. Verb choices vary semantically by operation, but the convention is predictable throughout.

    Tool Count3/5

    The 22-tool surface is heavy for a single MCP server, though it covers several related areas: mutual funds, financial calculators, and stock/index quotes. Each tool has a plausible role, but the count sits in the borderline-heavy 16–25 range.

    Completeness4/5

    Core mutual-fund workflows are well covered: search, scheme details, NAV history, SIP returns, portfolio valuation, comparison, rolling returns, tax estimation, and planning calculators. Minor gaps remain, such as richer scheme metadata like AUM/expense ratio and broader stock search/fundamental operations.

    Maintenance

    ActivityNo data
    ResponsivenessOutdated data

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      C
      maintenance
      Enables querying Indian mutual fund NAV data through MFAPI.in, providing net asset value information and fund details.
      99 npm
      MIT
    • A
      license
      Not graded
      quality
      A
      maintenance
      Enables AI agents to manage and analyze personal investment portfolios, including fund and stock holdings, net value tracking, XIRR calculations, penetration analysis, and backtesting.
      Apache 2.0
    • A
      license
      Not graded
      quality
      C
      maintenance
      Provides AI assistants with real-time Indian mutual fund NAV data from AMFI's official feed, requiring no API key.
      MIT
    • F
      license
      Not graded
      quality
      C
      maintenance
      Enables querying Indian mutual fund NAV analytics via DuckDB, offering tools for fund search, point-to-point returns, category rankings, and more, using local or Azure parquet data.
      -