How to Start Algo Trading: A Careful Path for India
To start algo trading in India, do the slow steps first: learn what it is, read the broker and exchange rules, write your trading rules in plain words, test them on past data, and practise with no real money. Only then decide whether to go live at all. Retail algo trading runs through a stock broker, and it does not make decisions better, only faster.
Risk note, up front: trading carries a real risk of losing money, and automation can lose it faster than you can react. This guide is education, not investment advice. It names no strategy, signal, broker or expected return.
What do I need before I start?
| What you need | Why it matters | How to check it |
|---|---|---|
| A trading rule you can write in plain words | A program can only follow rules that are explicit | Can someone else read your rule and say what it does? |
| A stock broker that allows API access to retail clients | Retail algo orders go through your broker | Ask the broker in writing |
| An amount you can afford to lose entirely | Any live system can lose money | Decide the number before you write any code |
| Time to test | Backtesting and paper trading take weeks, not hours | Plan the stages in advance |
| A way to stop everything | A bug or a fast market needs a hard stop | Know where the kill switch is |
What is algo trading, in one sentence?
Algo trading is trading in which orders are generated by automated rules instead of typed in by hand. Our guide to what algo trading is covers the types and the risks. Read it first if a term is new.
Step 1: Which rules apply to me in India?
Retail algo trading is allowed through a stock broker, within rules set by SEBI (the securities regulator) and the stock exchanges. Our guide on whether algo trading is legal in India walks through the details and the sources. From that guide, the points that matter most to a beginner are:
- API access comes from your broker, and a client using API access must give the broker a static IP address (a fixed internet address that does not change).
- Orders sent by algorithms are tagged, and there is a threshold, initially set at no more than 10 orders per second per exchange, above which an algorithm must be registered with the exchange through the broker.
- SEBI extended the implementation timeline in a circular dated 30 September 2025, so a blog post is never the final word.
Confirm the current text in the SEBI circulars and NSE documents listed under Sources, or ask your broker.
Step 2: What should I ask my broker?
Ask in writing, and keep the reply.
- Do you offer API access to retail clients, and what do you need from me?
- Do I need a static IP for my setup?
- How is my strategy classified under the current rules?
- Which order types are allowed through the API?
- What happens if my program sends an order outside the limits?
If you plan to use a third-party tool or vendor, also ask whether the vendor is empanelled with the exchange. Our note on what to check on a broker API helps here.
Step 3: How do I write my trading rules down?
Write every rule in plain words before any code exists. A rule you cannot write down cannot be tested. Use a template like this one, where every bracket is yours to fill in:
- Instrument: [what you will trade]
- Entry condition: [the exact event that triggers an order]
- Exit condition: [the exact event that closes it]
- Maximum order size: [a number]
- Maximum loss for the day: [a number]
- Stop-everything condition: [what makes the program halt]
This is a template, not a strategy.
Here is a worked example of the loss limit, with invented figures for illustration only. Suppose you decide you can afford to lose ₹10,000 in total on this experiment, and you set the daily loss limit at 5% of that amount. The limit is then ₹500 a day. The point is that you chose the limit before money was at risk.
Step 4: How do I test the rules on past data?
A backtest is a test of your rules against historical prices. It shows how the logic would have behaved in the past, not what it will do next, and a rule tuned too closely to old data can fail on new data. Our guide on how to backtest a trading algorithm explains the steps and the traps. A good backtest is a reason to keep testing, not to go live.
Step 5: What is paper trading, and why do it?
Paper trading means running your rules on live prices with no real money, so you see how the program behaves in real time: delays, rejected orders and bugs. Results still differ from real trading. Read paper trading vs live trading for the differences.
Step 6: Should I go live at all?
That is your decision, and it is fine for the answer to be no. If you do, first set the limits from Step 3 inside the program and at your broker, keep a log of every order the program sends, and protect your API keys. Decide in advance the smallest size you would still be comfortable losing completely.
Do I need to know how to code?
Not always, but you must understand what the system does. Our guide on what to learn before you code in Python is a starting point if you code. If you work with a developer, the note on what to write down before hiring an algo trading developer helps you prepare. A developer can make your rules explicit, test them and run them within your broker's limits. A developer cannot make a strategy profitable. The services on our home page describe what we build.
What mistakes should a beginner avoid?
- Trusting promised returns. In September 2022 SEBI issued a circular on performance and returns claimed by unregulated platforms offering algorithmic strategies (SEBI/HO/MIRSD/DOP/P/CIR/2022/117). Treat any promised return as a red flag.
- Sharing your login or API keys. This is practical security advice. Keep keys private and out of shared code.
- Skipping the test stages. Going straight to live trading lets a small bug become a real loss.
- Having no stop. Decide the maximum loss and the kill switch first.
Frequently asked questions
Is algo trading legal for retail investors in India?
Yes, through a stock broker and within SEBI and exchange rules. The rules are specific and have changed more than once, so confirm the current ones with your broker and the official documents.
How much money do I need to start algo trading?
We cannot give a reliable figure, and it depends on your broker and what you trade. Decide an amount you can afford to lose entirely, and check your broker's requirements.
Can I start algo trading without coding?
Possibly, using a tool or a developer. You still need to understand the rules the system follows, because you are responsible for the orders it places.
Is algo trading a way to make guaranteed profits?
No. It automates rules; it does not make them profitable. Any claim of guaranteed or typical returns should be treated with suspicion.
What this guide does not cover
It does not recommend a strategy, broker, platform or course, give return figures, or replace legal or financial advice. Costs are unavailable here because they differ by broker. Rules and dates change, so confirm them against the sources below.
Sources
- SEBI, circular SEBI/HO/MIRSD/MIRSD-PoD/P/CIR/2025/132, "Extension of timeline for implementation of SEBI Circular dated February 04, 2025 on 'Safer participation of retail investors in Algorithmic trading'", 30 September 2025: https://www.sebi.gov.in/legal/circulars/sep-2025/extension-of-timeline-for-implementation-of-sebi-circular-dated-february-04-2025-on-safer-participation-of-retail-investors-in-algorithmic-trading-_96979.html
- NSE, circular NSE/INVG/67858, "Safer participation of retail investors in Algorithmic trading", 5 May 2025: https://nsearchives.nseindia.com/content/circulars/INVG67858.pdf
- NSE, "Frequently Asked Questions (FAQs): Safer participation of Retail investors in Algorithmic trading": https://nsearchives.nseindia.com/web/sites/default/files/inline-files/FAQ_Retail Algo_03112025_NSE.pdf
- SEBI, circular SEBI/HO/MIRSD/DOP/P/CIR/2022/117, 2 September 2022: https://www.sebi.gov.in/legal/circulars/sep-2022/performance-return-claimed-by-unregulated-platforms-offering-algorithmic-strategies-for-trading_62628.html
This article is general information, not legal or financial advice. Trading carries risk.