GoldenPi Bonds MCP Server
Provides tools to list and search bonds from GoldenPi via reverse-engineered API endpoints.
README
GoldenPi Bonds MCP Server
A Model Context Protocol server that exposes GoldenPi's reverse-engineered bond discovery APIs.
Tools:
goldenpi_list_bonds— filtered, paginated bond listing viaPOST /v0/bonds/listgoldenpi_search_bonds— free-text bond search viaPOST /v0/bonds/search
Installation
cd goldenpi-mcp
python -m venv .venv
source .venv/bin/activate
pip install -e ".[dev]"
cp .env.example .env
Configuration
Copy .env.example to .env and edit as needed:
cp .env.example .env
Key variables:
| Variable | Default | Purpose |
|---|---|---|
GOLDENPI_BASE_URL |
https://api.goldenpi.com |
GoldenPi API base URL |
GOLDENPI_CLIENT_TOKEN |
(none) | Optional static token. If unset or rejected, the server falls back to /v0/auth/handshake. |
GOLDENPI_PARTNER_ID |
GoldenPi-Securities |
Partner ID sent during handshake |
GOLDENPI_SEND_DEPOSITORY_HEADER |
0 |
Set to 1 to send x-gpi-client-depositoryid on POSTs. Off by default because the live API rejects it on bond endpoints. |
GOLDENPI_MCP_TRANSPORT |
sse |
sse for ChatGPT apps / Render, stdio for OpenCode/Claude Desktop |
GOLDENPI_MCP_HOST / GOLDENPI_MCP_PORT |
0.0.0.0 / 8080 |
SSE server bind address. The standard PORT env var takes precedence over GOLDENPI_MCP_PORT. |
Auth behavior:
- Uses
GOLDENPI_CLIENT_TOKENif set. - If the static token fails with an auth/header error, automatically calls
POST /v0/auth/handshaketo obtain a guest token and retries. - If handshake also fails, the server exits immediately with a clear error.
Note: As of the latest test, the embedded static token from the frontend bundle is rejected by the live API. The handshake guest token works, so leaving GOLDENPI_CLIENT_TOKEN unset (or commented out) is the most reliable default.
Header note: The frontend only sends x-gpi-client-token on POST requests. Sending x-gpi-client-depositoryid on POST /v0/bonds/list causes an "invalid header" error, so this server does not send it by default. Set GOLDENPI_SEND_DEPOSITORY_HEADER=1 only if you specifically need it.
Running for ChatGPT / SSE clients (default)
python -m goldenpi_mcp.server
By default the server runs in SSE mode on 0.0.0.0:8080. Your ChatGPT app should connect to:
http://localhost:8080/sse
On Render, set the start command to python -m goldenpi_mcp.server and the server will automatically bind to the PORT env var Render provides.
Running locally (stdio / OpenCode / Claude Desktop)
python -m goldenpi_mcp.server --transport stdio
OpenCode / Claude Desktop config:
{
"mcpServers": {
"goldenpi-bonds": {
"command": "python",
"args": ["-m", "goldenpi_mcp.server", "--transport", "stdio"],
"env": {
"GOLDENPI_DEPOSITORY_ID": "GoldenPi"
}
}
}
}
Tool reference
goldenpi_list_bonds
Filtered, paginated bond discovery via POST /v0/bonds/list. Does not support free-text search — use goldenpi_search_bonds for name/ISIN/issuer queries.
Pagination & sorting
| Parameter | Type | Description |
|---|---|---|
sort_by |
enum | How to sort. Default: yield-high-to-low. |
limit |
int | Results per page. Range: 1–50. Default: 10. |
offset |
int | 1-based page number. Default: 1. |
Core filters
| Parameter | Type | Description |
|---|---|---|
asset_classes |
list | Instrument types: NCD, NCD-IPO, GSEC, SDL, CGTB, SGB-IPO. Defaults to NCD, NCD-IPO. |
credit_ratings |
list | aaa, aa, a, sovereign. sovereign matches government securities. |
min_yield / max_yield |
number | YTMC % range. Omit either side for open-ended. |
min_investment_inr / max_investment_inr |
number | Settlement amount per lot in INR. |
payment_frequencies |
list | annual, half-yearly, quarterly, monthly, thrice-yearly, on-maturity. |
issuer_types |
list | nbfc, corporate, psu, psb, bank, government, gsec, sdl, rbi. gsec/sdl also enable government asset classes. |
min_tenure_months / max_tenure_months |
int | Remaining tenure in whole months. Ignored if tenure_preset is set. |
tenure_preset |
enum | Shortcut: less-than-1y, 1-5y, 5-10y, more-than-10y, short-term-3y. |
investment_purpose |
enum | Strategy preset applying multiple filters at once, e.g. highest-safety, tax-free-investment, invest-short-time. |
Boolean flags
| Parameter | Description |
|---|---|
tax_free_only |
Tax-free bonds only. Mutually exclusive with taxable_only. |
taxable_only |
Taxable bonds only. |
nri_eligible_only |
NRI-eligible bonds only. |
secured_only / unsecured_only |
Secured or unsecured only. Mutually exclusive. |
perpetual_only / non_perpetual_only |
Perpetual or non-perpetual only. Mutually exclusive. |
call_option_only / no_call_option_only |
With or without call option. Mutually exclusive. |
pledge_for_fno_only |
Accepted as pledge for F&O margin. |
huf_tax_exemption_only |
HUF tax exemption available. |
spotlight_only |
Featured/spotlight bonds only. |
Advanced filters
| Parameter | Type | Description |
|---|---|---|
issuer_mode |
enum | public (Public Placement) or private (Private Placement). |
face_values |
list | 1000, 10000, 100000, 200000, 1000000, above-10-lacs. Combined into a range. |
bond_price_level |
enum | discount, par, premium. |
bank_bond_tier |
enum | tier_i, tier_ii, tier_iii. |
seniority |
enum | senior or subordinate. |
depository_listing |
enum | nsdl, cdsl, both. |
listed |
list | Override listing filter. [1] = active deals, [0] = primary/IPO, [0,1] = both. |
include_marketing_details |
bool | Include featured/UTSAV campaign metadata. |
Output:
{
"totalCount": 123,
"filteredCount": 45,
"hasMore": true,
"bonds": [ /* full instList rows */ ]
}
goldenpi_search_bonds
Free-text search across bonds/FDs/IPOs via POST /v0/bonds/search.
| Parameter | Required | Description |
|---|---|---|
search_filter |
yes | Name, ISIN, or issuer text. Minimum 2 characters. |
sort_by |
no | yield-high-to-low (default) or yield-low-to-high. |
limit |
no | Results per page. Range: 1–50. Default: 10. |
offset |
no | 1-based page number. Default: 1. |
Testing
pytest
Legal / risks
This is an unofficial reverse-engineered integration. GoldenPi may change their API without notice. Use at your own risk and respect their terms of service.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.