Live Demo · Data in 2 Minutes

Get free market data
with one prompt.

No coding. No spreadsheets. You type one sentence to Claude, paste two commands, and you have 1-month of minute-by-minute NIFTY & BANKNIFTY data.

Time: 2 minutes

You need: a free Dhan account

Cost: zero

What you'll have at the end

Real market data. On your screen.

7,894

NIFTY candles
(1 per minute)

7,894

BANKNIFTY candles
(1 per minute)

21

Trading days
(1 month of data)

Each candle tells you: what price did the market open at, what was the highest, the lowest, and where did it close — every single minute. That's what strategies are built on.

You don't need to understand the code. You just need to ask the right question.

Step 1 · The Prompt

You type one sentence.

What you type to Claude

“Download 1 month of 1-minute NIFTY and BANKNIFTY data from Dhan and save it.”

That's it. Claude figures out which library to use, writes the download script, handles errors and retries. You don't write a single line of code.

The skill is knowing what to ask for — not how to code it.

Step 2 · Claude gives you two commands

Copy. Paste. Done.

command 1

# Create a workspace
$ python3 -m venv data_venv

command 2

# Install what's needed
$ ./data_venv/bin/python -m pip install dhanhq pandas pyarrow

You don't need to understand these commands. Command 1 creates a clean workspace. Command 2 installs the tools Claude needs. You run them once and never again.

Two commands. One time. Then Claude does the rest.

Step 3 · Give Claude your Dhan credentials

Claude asks. You paste.

Claude asks

“What's your Dhan Client ID and Access Token?”

Client ID

A 10-digit number from your Dhan profile.

1111609313

Access Token

A long string from Dhan's API page. Expires daily.

eyJ0eXAiOiJKV1Qi...

Paste both. Claude downloads the data. Takes about 30 seconds.

Step 4 · Watch it download

30 seconds. 7,894 candles each.

downloading...

Downloading NIFTY (security_id=13)
  NIFTY 2026-04-20 → 2026-04-25 ... 1,877 rows
  NIFTY 2026-04-26 → 2026-05-01 ... 1,510 rows
  NIFTY 2026-05-02 → 2026-05-07 ... 1,502 rows
  NIFTY 2026-05-08 → 2026-05-13 ... 1,502 rows
  NIFTY 2026-05-14 → 2026-05-19 ... 1,503 rows
  Saved 7,894 rows → data/nifty_1min.parquet

Downloading BANKNIFTY (security_id=25)
  BANKNIFTY 2026-04-20 → 2026-04-25 ... 1,877 rows
  ...
  Saved 7,894 rows → data/banknifty_1min.parquet

Done.

Two files. ~8,000 candles each. Every minute of every trading day for a month.

Step 5 · See your data

Ask Claude: “show me the data in my browser.”

What you type

“Show me the data in my browser.”

Claude converts the data into an HTML table and opens it. You see every candle — scrollable, browsable, on your screen.

browser preview

NIFTY — 1 Minute Candles
7,894 candles · 2026-04-20 to 2026-05-19 · 21 trading days

timestamp             open       high       low        close      volume
2026-04-20 09:15:00   24,391.50  24,420.20  24,270.20  24,295.85  8,087,933
2026-04-20 09:16:00   24,294.35  24,314.15  24,241.25  24,269.40  5,201,414
2026-04-20 09:17:00   24,270.10  24,322.25  24,264.45  24,318.25  3,471,314
2026-04-20 09:18:00   24,320.50  24,325.95  24,291.35  24,304.10  3,865,764
2026-04-20 09:19:00   24,301.35  24,339.55  24,299.60  24,339.55  2,348,044
... 7,889 more rows

Real prices. Real volumes. Every minute. Now the fun part — asking questions.

Step 6 · Now explore

Ask Claude anything about your data.

“Which 1-minute candle had the biggest high-to-low swing?”

Finds the most volatile minute in a month. Spoiler: it's always 09:15.

“How many days did the market close higher than it opened?”

Instant green-vs-red day count. This month BANKNIFTY was red 71% of days.

“What time of day has the highest average volume?”

Maps out when the market is busiest. Useful for timing entries.

“Show me a chart of NIFTY closing prices.”

Claude plots it. You see the trend without opening any charting software.

“If I bought at 09:20 and sold at 15:00 every day, what would my P&L be?”

Your first backtest. One sentence. Claude does the math.

“What's the average range of the first 15 minutes vs the last 15 minutes?”

Compares opening volatility to closing volatility. Real trading insight.

Behind the scenes

What Claude did (so you didn't have to)

1

Wrote the script

Picked the right library (dhanhq), figured out the API call, handled the response format. ~50 lines of Python.

2

Debugged 3 problems

The library changed its import name. The data came in an unexpected format. The API silently rate-limited. Claude fixed all three.

3

Converted timestamps

Raw data had Unix epoch numbers. Claude converted them to readable dates in IST (Indian Standard Time).

4

Made it browsable

Parquet files are binary — can't open in a browser. Claude generated a styled HTML table you can scroll through.

The code is simple. The debugging is where the time goes. That's what AI handles for you.

The real lesson

You don't need to
memorize APIs.

The skill is knowing what data exists, what questions to ask, and how to describe what you want. Claude handles the plumbing.

End of Demo

One prompt. Real data.
Infinite questions.

Next: Use this data to build your first backtest →