The scope is deliberately small: historical candles and nothing else — no quotes, no orders, no positions. Symbol resolution comes from a CSV in the repo rather than a live lookup, so a ticker that is missing there is missing for you too; check the file before assuming a symbol is unsupported.
A market-data server for India's National Stock Exchange, reading through the Upstox API. It answers with historical candles, so an assistant can reason about a price series instead of guessing at one.
- `get_historical_candle` returns open, high, low, close and volume for an NSE symbol over a date range you choose
- Five intervals: `minute`, `30minute`, `day`, `week` and `month`
- Validation and error messages that name the problem — an unknown symbol, a bad date format, a connection failure or an upstream API error
- Symbols and their instrument keys come from `configs/NSE.csv` in the repository
The project nse-ticker-mcp, version 0.1.0, installed from a checkout: clone it, make a virtualenv, `pip install -e .`. Python 3.11 or higher, with mcp, pandas, httpx and pydantic pulled in as dependencies. The client launches `ticker.py` from the project directory.