Free · No API key · Open source

BondDekho MCP Server

Connect Claude, ChatGPT, Cursor, or any MCP-compatible AI assistant to live Indian corporate bond data. Search 3000+ bonds, compare yields across 9+ platforms, and ask questions in plain English.

For Developers

BondDekho ships a Model Context Protocol server in two flavors: a hosted SSE endpoint (zero install) and a local stdio package from PyPI. Pick whichever fits your MCP client.

1

Remote (no install)

Recommended

Point your MCP client at the hosted SSE endpoint. Works in Claude Desktop, Cursor, Windsurf, and any MCP client supporting SSE transport.

~/Library/Application Support/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "bonddekho": {
      "url": "https://www.bonddekho.com/mcp/sse"
    }
  }
}
2

Install via pip (stdio)

Runs locally over stdio transport. Useful if your environment blocks SSE or you want to pin a specific version.

# Install
pip install bonddekho-mcp

# Then add to your MCP client config:
{
  "mcpServers": {
    "bonddekho": {
      "command": "bonddekho-mcp"
    }
  }
}

Package: pypi.org/project/bonddekho-mcp

3

Claude Code

Drop a .mcp.json in your project root.

# In your project root, create .mcp.json:
{
  "mcpServers": {
    "bonddekho": {
      "url": "https://www.bonddekho.com/mcp/sse"
    }
  }
}

SSE Transport

Real-time streaming over HTTP. No long-poll, no WebSocket setup.

Read-only

DB opened with mode=ro. No tool can mutate state.

MIT License

Source on GitHub. Fork, audit, self-host — your call.

For Investors

If you already use Claude, ChatGPT, or another AI assistant, you can ask it about Indian bonds and get live data — not a year-old training-data answer.

What is MCP?

Model Context Protocol (MCP) is an open standard from Anthropic that lets AI assistants connect to external data sources. Once you add BondDekho as an MCP server in your AI app, the assistant can look up live bond yields, compare bonds across platforms, and pull market stats — all without you leaving the chat.

Think of it as giving your AI assistant a direct line to BondDekho's database, the same one that powers the website.

Example questions you can ask

  • What bonds does Muthoot Finance offer?
  • Show me AAA rated bonds with YTM above 8%
  • Compare bonds INE101Q07AS3 and INE296G07218
  • What is the average bond YTM across India's platforms?
  • Find NBFC bonds available on GoldenPi under Rs. 10,000 minimum

A note on safety

The MCP server only reads data — it cannot transact, place orders, or move money. Every response includes a disclaimer that BondDekho is not a SEBI-registered adviser. All information is educational. Always verify with the original platform and consult a SEBI-registered adviser before investing.

Available Tools

Eight tools are exposed to MCP clients — five data lookups and three calculators. The assistant picks the right one based on your question.

search_bonds(query, rating?, min_ytm?, max_ytm?, platform?, limit?)

Search bonds by issuer name, ISIN, or sector. Filter by rating, YTM range, or platform.

get_bond(isin)

Get full details for a single bond including platform-by-platform pricing.

compare_bonds(isins[])

Side-by-side comparison of 2-5 bonds in a markdown table.

get_market_stats()

Overall market statistics: total bonds, average YTM, rating distribution, per-platform overview.

get_issuer_bonds(issuer)

All bonds from a specific issuer with platform data, sorted by YTM.

calc_ytm(current_price, coupon_rate, years_to_maturity, face_value?, frequency?)

Calculate yield to maturity given the bond’s market price and cash flow profile.

calc_bond_price(coupon_rate, ytm, years_to_maturity, face_value?, frequency?)

Calculate the fair price of a bond for a target YTM.

calc_post_tax_return(investment_amount, ytm, holding_period_years, tax_slab_percent, listed?)

Compute post-tax effective yield in India (LTCG 12.5% for listed bonds, slab rate otherwise).

Available Resources

Resources are static context that MCP clients can pre-fetch (no tool call needed). Available on the remote SSE endpoint only.

bonddekho://blog/index

Index of all 40+ BondDekho blog posts and weekly market reports.

bonddekho://blog/{slug}

Full markdown content of a single blog post.

bonddekho://glossary/index

Alphabetised index of all 39 bond terminology entries.

bonddekho://glossary/{slug}

Full definition of a single glossary term (with related calculator link).

Frequently Asked Questions

Do I need an API key?

No. The remote SSE endpoint at https://www.bonddekho.com/mcp/sse is open and rate-limited per IP. The pip package runs entirely locally and reads BondDekho's public data.

How fresh is the data?

Bond data is refreshed 4 times daily (6 AM, 12 PM, 6 PM, 11 PM IST). The MCP server reads from the same SQLite database that powers bonddekho.com.

Which AI clients are supported?

Anything that speaks MCP: Claude Desktop, Claude Code, Cursor, Windsurf, Cline, Continue, and others. The remote endpoint uses SSE; the pip package uses stdio — most clients support one or both.

Can the MCP server place trades or move money?

No. The server is strictly read-only. The database connection is opened in read-only mode (mode=ro). There are no write tools, no transaction endpoints, no order placement. You still go to GoldenPi / WintWealth / IndiaBonds / etc. to actually buy a bond.

Is the source code available?

Yes — MIT licensed. The pip package bonddekho-mcp ships with the same server.py that powers the remote endpoint.

Important Disclaimer

BondDekho is a data aggregation platform, not a SEBI-registered investment adviser. The MCP server provides information for educational purposes only. Do your own research and consult a SEBI-registered adviser before investing. Bond investments carry credit, interest rate, and liquidity risks. See legal disclosures for details.

MCP Server — Query Indian Bond Data from AI Assistants | BondDekho