← AITrading.in · Blog

Is Algo Trading Legal in India? A Plain-English Guide

Short answer: yes. Algorithmic trading is allowed in India for retail investors, but only through your stock broker and under rules set by SEBI and the stock exchanges. The rules are specific, and they have changed more than once, so check them before you automate anything that touches real money.

This guide was checked against the SEBI and NSE documents listed at the end, read on 17 October 2026. It does not claim to cover every circular issued after that date.

What the rules cover

SEBI's circular of 4 February 2025 set up the retail framework. NSE's implementation standards, circular NSE/INVG/67858 dated 5 May 2025, describe the automated trading it covers as "automated trading systems, computer algorithms, or other software-based trading solutions to automate the buying and selling of securities/contracts on the secondary markets" (section B.1).

Your broker is the gate

Retail API access comes from your broker. NSE's standards say a client using API access "must mandatorily provide the stockbroker with a static IP address(es)" (sections A.1 and A.5). They also say brokers must not allow open APIs. Access has to go through a client-specific API key and a static IP the broker has whitelisted, with OAuth-based authentication and two-factor authentication (sections I.c to I.e). A broker is "fully responsible and liable for all orders emanating through" its client API systems (section I.f).

Whatever your program does, its orders pass through your broker's risk controls.

Static IP, daily logout and the 10-orders-per-second line

Three rules matter for most retail setups:

Below the threshold, the client does not have to register the algo with the broker's system. Orders still need an exchange-provided tag, because all algo orders, below and above the threshold, must be tagged (section G). Above the threshold, the client must register each algorithm with each exchange where it will be used, through the broker, and orders are tagged with the exchange-issued algo ID (sections C.1 and C.2). Brokers must reject orders that exceed the limit (section B.5).

Where the strategy runs

NSE's FAQ on the retail framework quotes NSE/INVG/69255 of 22 July 2025, paragraph 14: "all the strategies shall be run on the brokers servers. The order messages shall be originated from brokers server." The same FAQ says a tech-savvy client who writes their own algo can host it at their own static IP instead. It also says a client static IP is required only for a tech-savvy investor using the API (questions 3, 5 and 6).

Ask your broker in writing how your setup is classified.

Algo vendors must be empanelled

NSE's standards say algo providers must be empanelled with the exchanges and registered with each exchange where their algorithms trade. Brokers must tell the exchanges about any commercial or technical arrangement with an empanelled provider, and must carry out due diligence on that provider (sections E.1, E.4 and E.5). If someone offers to run a strategy for you, ask which empanelled provider they work with and which broker the orders go through.

For "black box" algos, where the logic is not disclosed, the FAQ quotes the SEBI circular as requiring the provider to register as a Research Analyst and keep a detailed research report for each algo (question 4).

The timeline moved, twice

SEBI's circular of 30 September 2025 (SEBI/HO/MIRSD/MIRSD-PoD/P/CIR/2025/132) says the framework was first due on 1 August 2025. A circular dated 29 July 2025 extended the start to 1 October 2025. Brokers that were not ready got a glide path with milestones on 31 October 2025, 30 November 2025 and 3 January 2026 (paragraphs 1 and 4). Brokers that missed those milestones were barred from onboarding new retail clients for API-based algo trading from 5 January 2026 (paragraph 5). From 1 April 2026, the framework and the exchanges' implementation standards apply to all stock brokers (paragraph 8).

What is not a safe shortcut

Legal in principle does not mean anything goes.

A practical checklist

  1. Ask your broker in writing: Do you offer API access to retail clients? What do you need from me, such as a static IP? How is my strategy classified?
  2. Test first. Run the logic on historical data, then in a no-money mode. A backtest shows how the logic would have behaved on past data. It does not predict what happens next.
  3. Build in limits. Set a maximum order size, a maximum loss for the day, and a kill switch you can reach from your phone.
  4. Keep logs and protect your keys. Record every order your program sends, and keep API keys out of shared code and public repositories.
  5. Re-check the rules. The dates above have moved more than once. Read your broker's notices and the current exchange circulars before you go live.

Where a custom build fits

Some traders write their own code. Others work with a developer on a system that connects to a broker API, tests on historical data, and provides options analytics. If you are in the second group, the compliance questions above should come up in the first conversation. The services section on the home page lists the kinds of systems built, options analytics is covered here, and the process from requirement to live system is here. The FAQ answers common questions.

A developer cannot make a strategy profitable. Software can make your rules explicit, test them on data, and run them consistently within your broker's and the exchange's limits.

Bottom line

Sources

This article is general information, not legal or financial advice.