What a custom trading bot costs, and what drives the price

The honest answer to "what does a trading bot cost?" is that the same seven questions decide it every time, and until they are answered any number is a guess. This article is the list. Read it and you will know, before you write to anyone, whether yours is a ten-day build or a three-month one, and what you can change to make it the former.

Why there is no price list

"A trading bot" describes a moving-average crossover on one stock and a multi-account options system that reconciles positions across three brokers. Both are bots. One is a few days of careful work; the other is a platform. A price list would either overcharge the first person or promise the second something that cannot be delivered for the money. So we do not publish one, and we are wary of anyone who does. What we do publish is what moves the number.

The seven drivers

Each of these has a lighter version and a heavier version. Where your project sits on each one is most of the quote.

1. The broker and its API

Lighter: a broker with a mature, documented API that we have built on many times. Interactive Brokers, Zerodha's Kite Connect, MetaTrader 5, TradeStation. Heavier: a broker with a thin or undocumented API, one that changes without notice, or one that needs a workaround such as browser automation. Sometimes the best decision for the budget is to open an account at a broker with a proper API.

2. How many instruments and markets

Lighter: one instrument type in one market: US stocks, or NSE futures, or one forex pair. Heavier: options chains (expiries, strikes, multi-leg orders), several markets with different hours and holidays, or a universe of hundreds of symbols that has to be scanned every bar. Options roughly double the order-handling work on their own.

3. Execution complexity

Lighter: market or limit orders, one position at a time. Heavier: bracket orders with trailing stops, pegged orders, orders that must fill together or not at all, partial-fill handling, and anything where the order logic itself is part of the strategy. Most of the engineering in a live system is here and in the next item, not in the entry signal.

4. Data

Lighter: your broker's own feed is enough. Daily bars, or intraday bars at the speed the broker delivers them. Heavier: a strategy whose rules only make sense on true tick data or on a consolidated feed, which means a paid vendor, a data-handling layer, and a bigger backtest. We will tell you which yours is. Our guide to market-data engineering covers how we size the feed to the strategy rather than the other way round.

5. Running unattended

Lighter: you start it in the morning, watch it, stop it in the evening. Heavier: it runs on a server on its own, so it needs reconnection without duplicate orders, reconciliation of its positions against the broker, a kill switch, alerts to your phone, and a clean restart after a crash. This is the difference between a demo and a system, and it is where a "simple" bot quietly becomes a real project. The edge cases guide lists what has to be handled.

6. Dashboards, accounts and brokers

Lighter: one account, one broker, a log file and a console. Heavier: a web dashboard, several accounts each with their own limits, or two brokers in one system. Every one of these adds a layer that must stay consistent with the others. This is the line between a starter build and a platform.

7. Ownership

Standard: a working system, we keep the source. Ownership: a full licence to the code, yours to keep and modify, for an additional 30% of the fee. Not a driver of the engineering, but a line on the quote, so it belongs here.

Two scopes, side by side

Take one strategy, say a rule on 15-minute bars for US stocks, and imagine it built two ways.

• Starter build. One broker with a good API, your broker's own data, market and limit orders, one position at a time, you watch it during the session, no dashboard. About ten days from an agreed scope. This is the starter trading bot.
• Platform build. The same rule, but on a server, unattended, across three accounts with per-account limits, bracket orders with trailing stops, a dashboard showing positions and P&L reconciled against the broker, alerts, and a second broker planned for next year. Months, not days, and most of the time goes into items 3, 5 and 6, not into the rule.

Same strategy, entirely different projects. That is why the brief matters more than the idea.

What you can do to keep it small

• Write the rules down before you write to anyone. Entry, exit, sizing, stops, in plain words. Every ambiguity we have to ask about is time; every one we have to guess at is risk.
• One broker, with an API subscription already active. Getting API access approved can take longer than the build.
• Use the broker's data unless the rules genuinely need more. Most positional and swing rules do not.
• Start attended. Run it while you watch for a few weeks. Add the unattended layer once you trust it. That order costs less than the reverse.
• Freeze the rules for the build. Ideas that arrive mid-build go on a list for version two. Changing rules while the code is being written is the most common way a ten-day build becomes a month.

What we will not quote for

A strategy. We build to your rules; we do not supply them, and we do not tune yours for results. "Make it profitable" is not a scope. Nor is a project without written rules: we can help you write them down in a call, but the rules stay yours. And nothing on this page or on this site is investment advice or a claim about what any system will earn. Trading involves substantial risk of loss.

How to get a real number

Send the six things: market and broker, rules in plain words, timeframe and instruments, data source, whether you want to own the code, and the budget you have in mind, in any currency. The starter trading bot page shows two example briefs, one for US markets and one for India. We reply with our questions, then a fixed quote, usually within two working days. Or just tell us what you want built and we will tell you which of the seven drivers apply.