PyZData – Indian Historical Market Data Downloader

Download historical OHLCV and Open Interest candle data for any stock, index, or F&O instrument from Zerodha — as a Python library, a CLI command, or a browser-based web app. No API subscription required.

What you can do with PyZData

• Download price history for any stock or index listed on Zerodha (NSE, BSE, NFO, MCX)
• Choose intervals from 1-minute to daily
• Export clean pandas DataFrames ready for backtesting, charting, or ML
• Use the web app — no coding needed
• Use the CLI — one-liner downloads from the terminal
• Use the Python library — import and integrate into your own scripts

Web Interface (no coding required)

The easiest way to use PyZData is the built-in Streamlit web app.

Run it:

pip install "pyzdata[web]"
pyzdata-web

Then open http://localhost:8501 in your browser.

Everything is on one page:

1. Log in — paste your enctoken (checked immediately) or use your user ID, password and TOTP
2. Pick an instrument — a searchable dropdown of everything Zerodha lists; for futures & options pick the underlying, then the contract
3. Pick a period and interval — one click each (1W … 5Y or custom dates; Day … 1 min)
4. Download — see a price chart and save the result as CSV or Excel (.xlsx)

Installation

# Library only (Python API + CLI)
pip install pyzdata

# Library + web interface (adds streamlit)
pip install "pyzdata[web]"

Requirements: Python 3.10+, pandas ≥ 2.3, requests ≥ 2.32

For development:

git clone https://github.com/vikassharma545/Historical-Market-data-From-Zerodha.git
cd Historical-Market-data-From-Zerodha
pip install -e ".[web,dev]"

Python Library

from pyzdata import PyZData, Interval

# Login with enctoken (paste from browser cookies after logging into kite.zerodha.com)
client = PyZData(enctoken="your_enctoken_here")

# OR login with credentials
client = PyZData(user_id="AB1234", password="your_password", totp="123456")

# Find the instrument token for a symbol
token = client.get_instrument_token("NIFTY 50", "NSE")

# Not sure of the exact symbol name? Search for it
results = client.search_instruments("NIFTY", exchange="NSE")
print(results[["tradingsymbol", "instrument_token", "exchange"]])

# Download daily candles for a full year
df = client.get_data(token, "2024-01-01", "2024-12-31", Interval.DAY)
print(df.head())

# Download 1-minute candles with Open Interest (F&O instruments)
fut_token = client.get_instrument_token("NIFTY24JANFUT", "NFO")
df = client.get_data(fut_token, "2024-01-02", "2024-01-25", Interval.MINUTE_1, oi=True)

Output columns: tradingsymbol, datetime, open, high, low, close, volume (+ open_interest when oi=True)

Symbol lookup forgives case and extra spaces ("reliance" works), and an invalid or expired enctoken is rejected as soon as the client is created.

Long ranges are split into as few requests as Zerodha allows — 5 years of daily candles is a single request; 1-minute candles need one request per 60 days.

Context manager — automatically closes the HTTP session when done:

with PyZData(enctoken="your_enctoken") as client:
    token = client.get_instrument_token("RELIANCE", "NSE")
    df = client.get_data(token, "2024-01-01", "2024-12-31", Interval.DAY)
    df.to_csv("reliance.csv", index=False)
# session is closed automatically here

CLI

# Daily data for NIFTY 50 — print to terminal
pyzdata download --enctoken TOKEN --symbol "NIFTY 50" --exchange NSE \
        --start 2024-01-01 --end 2024-12-31

# 1-minute data — save to CSV
pyzdata download --enctoken TOKEN --symbol RELIANCE --exchange NSE \
        --start 2024-01-01 --end 2024-06-30 \
        --interval minute --output reliance.csv

# Login with credentials
pyzdata download --user-id AB1234 --password pw --totp 123456 \
        --symbol RELIANCE --exchange NSE \
        --start 2024-01-01 --end 2024-01-31

# Search for a symbol
pyzdata search --enctoken TOKEN --query HDFC --exchange NSE

# Keep the token out of your shell history
export PYZDATA_ENCTOKEN=TOKEN        # PowerShell: $env:PYZDATA_ENCTOKEN = "TOKEN"
pyzdata download --symbol RELIANCE --exchange NSE --start 2024-01-01 --end 2024-12-31

# All options
pyzdata download --help
pyzdata search --help

Available Intervals

Configuration

Set any of these environment variables to override the defaults (.env.example lists them all):

Or pass a Config object in code:

from pyzdata import Config, PyZData

cfg = Config(max_workers=8, log_level="DEBUG")
client = PyZData(enctoken="...", config=cfg)

Error Handling

from pyzdata.exceptions import (
    AuthenticationError,    # wrong credentials or expired enctoken
    InstrumentNotFoundError, # symbol not found
    DataFetchError,          # API or network failure
    PartialDataError,        # some date ranges failed — carries the rest
    PyZDataError,            # catch-all base class
)

try:
    client = PyZData(enctoken="...")
    token  = client.get_instrument_token("RELIANCE", "NSE")
    df     = client.get_data(token, "2024-01-01", "2024-12-31", Interval.DAY)
except AuthenticationError as e:
    print(f"Login failed: {e}")
except InstrumentNotFoundError as e:
    print(f"Symbol not found: {e}")
except PartialDataError as e:          # must come before DataFetchError
    print(f"Gaps at {e.failed_ranges}")
    df = e.partial_data                 # everything that did download
except DataFetchError as e:
    print(f"Download failed: {e}")

PyZData never hands you data with silent gaps: a date range that still fails after a retry raises PartialDataError instead.

Project Structure

pyzdata/
├── client.py        PyZData — main entry point
├── auth.py          Two-step Zerodha login
├── instruments.py   Symbol lookup with 24-hour disk cache
├── downloader.py    Parallel, interval-aware data fetching
├── models.py        Interval enum
├── config.py        Settings + environment variable loading
├── exceptions.py    Typed exception hierarchy
├── cli.py           pyzdata command-line tool
├── _app.py          Streamlit web interface
├── _web_helpers.py  Streamlit-free logic behind the web interface
├── _xlsx.py         Fast, dependency-free Excel (.xlsx) writer
├── run_web.py       Entry point for pyzdata-web command
└── py.typed         PEP 561 type-checking marker

app.py               Local dev launcher for Streamlit
tests/               Unit tests (no credentials or network needed)
CONTRIBUTING.md      How to contribute
CHANGELOG.md         Release history
SECURITY.md          Security policy

Running Tests

pip install -e ".[web,dev]"
pytest

No Zerodha account or internet connection needed — all HTTP calls are mocked.

How to get your enctoken

1. Log in to kite.zerodha.com
2. Press F12 → Application tab → Cookies → kite.zerodha.com
3. Copy the value of the enctoken cookie
4. Paste it into the app or pass it as PyZData(enctoken="...")

The enctoken refreshes each time you log in to Kite.

License

MIT — see LICENSE

Contributing

Contributions are welcome! See CONTRIBUTING.md for guidelines.

Changelog

See CHANGELOG.md for the full release history.

Security

To report a vulnerability, see SECURITY.md.

Author

Built by Vikas Sharma