abtrade
Historical NSE data + Python sandbox · MCP

Real Indian-market data, inside your AI agent.

abtrade gives any MCP client — Claude Code, Cursor, your own agent — a direct line to real NSE price history and a Python sandbox. Generate a key, connect it, and just ask.

claude
$ claude
 what has NIFTY done this week?

abtrade · get_prices(["NIFTY"], days=5)
    NIFTY 50   24,219.05   ▼ 0.05%
    5d range   24,14424,313

  Roughly flat on the week — a 0.7% band, no trend.
Every figure is pulled from the database, live.
01
Real NSE data
Historical daily and intraday OHLCV across the NSE universe, served straight from the database — not scraped at request time, not approximated.
02
A Python sandbox
numpy / scipy / pandas over the full data API. Screens, correlations, option greeks, quick backtests — all as code your agent writes and runs.
03
Ask in plain English
Two tools — get_prices and analyze. Pose a market question; your agent composes the code, runs it on the data, and hands back the answer.
04
Every number is real
Results come from the database, never the model's imagination. If it can't be computed from the data, your agent says so instead of guessing.
Point your agent at real market data.