Tutorial · Strategy Lab

Create a trading bot with AI. No code.

Describe your strategy in one sentence. Sena writes the rules, Strategy Lab backtests them on two years of candles after costs, and one click turns them into an agent that trades a paper account of its own on live prices.

Education, not investment advice. Sena never places trades.

By Konsta · Updated November 2026

Step by step

Your first bot, in seven steps.

About five minutes, start to finish. Everything below is the real Strategy Lab, with a QQQ idea that held up over two years.

  1. 1Step 1 of 7

    Describe your strategy

    Open Strategy Lab with the flask button in the chart toolbar. In Write, say the symbol, the entry, the stop, the target and when to be flat, then press Build rules. Sena lists every assumption it made.

    Buy QQQ when a 5-minute candle closes above the first 30-minute range, only in a daily uptrend. Stop on the other side of the range, target the next level. One trade a day, flat by 3:55.

    Strategy Lab Write tab with a QQQ opening range idea, and what Sena assumed: 5-minute candles, an uptrend read as yesterday’s close above a rising 20-day average, the stop at the other side of the range, the next resistance as the target with 2R if none, and a first signal at 10:00

  2. 2Step 2 of 7

    Check every rule

    Each rule is a dropdown or a number: symbol and timeframe, direction, level, entry, filters, stop, target, trading hours and how much history to test. Fix anything Sena guessed wrong.
    Strategy Lab rules: QQQ on 5m, long only, opening range 30 min, entry on a close above, only when the daily trend is up against the 20-day average, stop beyond the level, target the next level with 2R if none, one trade a day, entries 10:00 AM to 2:00 PM, exit 3:55 PM, 24 months of history, and the Backtest and Turn on as agent buttons

  3. 3Step 3 of 7

    Backtest two years, after costs

    Pick up to 24 months and press Backtest. This one held up: +0.09R a trade over 149 trades, 60% winners, and Proven Edge above zero. It was the third idea we tried; the first two did not, and most ideas won’t.
    Strategy Lab Test tab: QQQ 5-minute, Dec 2 to Nov 6, 485 sessions after costs, verdict Edge, works with an edge: +0.09R a trade, 60% win rate, 149 trades, Proven Edge +0.01R, profit factor 1.32, max drawdown −5.4R, best entries in the first hour

  4. 4Step 4 of 7

    Turn it on as an agent

    Press Turn on as agent. Name it, give it its own paper account and set the risk per trade. It sizes every trade from the distance to its stop.
    Turn on as agent: it runs the rules on every 5-minute candle close in regular hours and trades a paper account of its own; name QQQ first 30-minute range, a $10,000 paper account and 1% risk per trade, or $100

  5. 5Step 5 of 7

    Set its limits and switch it on

    Choose when it stops for the day and how it tells you about trades. The agent is judged against the backtest you just ran, so you can see if it keeps up.
    Agent settings: stop for the day after 3R, paper trades always on, a toast with sound in the app, phone push and email coming later, judged against the backtest of 149 trades over 24 months at +0.09R a trade, and the Turn on agent button

  6. 6Step 6 of 7

    Watch it in the Agents tab

    Open Trading in the chart toolbar, then Agents. Each agent shows whether it is running, today’s result and its paper equity. Click it for win rate, drawdown and paper results next to the backtest’s range, or to show its trades on the chart, edit it or delete it.
    Trading panel, Agents tab: the QQQ first 30-minute range agent, v1, on 5-minute candles, switched on and running, waiting for a trade, today $0.00, paper $0.00, equity $10,000.00, 1 of 5 agents on Pro

  7. 7Step 7 of 7 · Example

    Judge it against its backtest

    Click the agent to see its paper P&L, win rate, drawdown and every trade, with its equity line drawn over the range its backtest expects. This example fills the view with 20 trades from the agent’s own backtest, its best stretch (Sep to Nov 2025): 16 winners and +8.3R, or +$830 at $100 a trade, above the typical range. A live agent fills it with its paper trades.
    Example agent detail filled with 20 trades from the QQQ backtest: paper P&L +$830.05, equity $10,830.05, 20 trades, 80% win rate, +0.4R a trade, max drawdown −$101, the paper equity line above the backtest range, and the trade list

Ways to start

Your idea, a pattern or a remix.

Eight doors in. Every one leads through a backtest, the bar the agent has to clear.

In your wordsOne sentence
A ready-made ideaOne click
Rules by handDropdowns
The Agents tabNew rule agent
Remix an agentSave as v2

Coming soon

Claude, ChatGPT, CursorMCP
Your chart alertsWebhook
Your own codeAPI key

Backtest

Up to 24 months, after costs

Works, with an edge
Agent

Running on paper

Equity

$10,000

Risk

1% a trade

Strategy Lab Write tab with a box to describe a strategy and three ready-made ideas: an opening range break long or short, a support bounce in a daily uptrend, and a VWAP rejection short with RSI above 60
Start from a ready-made idea
Editing the QQQ first 30-minute range agent: it keeps running on v1 until you save; saving makes v2 and starts a fresh paper comparison, earlier trades stay with v1, with Save as v2 and Cancel editing buttons
Remix a running agent. It keeps trading v1 until you save.

What every agent does

It trades your rules while you watch the market.

Pause it, edit it or delete it from one menu, any time.

Agents tab: the QQQ first 30-minute range agent switched on and running, 1 of 5 on Pro, with its menu open: Pause, no new entries until you resume; Edit in Strategy Lab, saving makes version 2 and paper history stays with version 1; and Delete

Every candle close

5m, 15m or 1h

Sized from its stop

The risk % you set

Own paper account

$5k to $50k

Daily loss stop

In R

Toast with sound

Every entry and exit

Judged vs backtest

On track · Watch · Off track

Pause your way

Keep or close the trade

Versioned

Old trades stay with v1

On your chart

Show on chart

Never real money

5 agents on Pro

Coming soon

Run your agents from Claude or ChatGPT.

Senntric is adding an MCP connector. Add Senntric in Claude, ChatGPT or Cursor, sign in with your Senntric account, and ask about your setups and agents in plain English, or tell an agent what to do.

Read

Your setups, positions and agent results.

Paper trade

Place paper trades through the same paper broker as the app.

Manage agents

Create, pause and stop agents. Turning one on asks you first.

Paper accounts only, with the same Pro limits as the app. Your AI can never place a real-money order.

Your AI assistant · example

How is my QQQ agent doing?

Senntric · agent_performance

QQQ first 30-minute range: 20 paper trades, 16 winners, +8.3R (+$830). That is above the range its backtest expects.

Pause it before CPI tomorrow.

Senntric · pause_agent

Ready to pause QQQ first 30-minute range. Confirm in Senntric.

What you can build

The building blocks Sena writes for you.

Levels

Opening range (5, 15 or 30 minutes), support/resistance zones, prior day’s high, low or close, premarket high or low, VWAP, or a moving average (SMA or EMA).

Entries

Break, touch, bounce or reclaim of the level, long, short or both.

Candle confirmation

None, engulfing, hammer or shooting star, or inside bar on the signal candle.

Conditions

Up to four filters: daily or intraday trend, RSI above or below a value, gap up or down, relative volume, and higher or lower volatility than usual.

Stops

Beyond the level or the signal candle (with an optional ATR buffer), an ATR multiple, or a percent of entry.

Targets

An R multiple, the next level, a trailing ATR stop, or no target with a time exit. Optionally take half off at 1R and move the stop to entry.

Timing

5-minute, 15-minute or 1-hour candles; an entry window and a flat time in New York time; up to five trades a day.

Calendar

Trade or skip high-impact US economic release days and the symbol’s earnings days.

Reading the result

Four verdicts, in plain English.

Results are in R: each trade’s result divided by its risk to the stop.

Edge

Works, with an edge

Proven Edge above zero: the profit is unlikely to be luck alone.

Unclear

Made money, not proven yet

Positive on average, but not by enough to rule out luck.

Unclear

Too few trades

Fewer than 30 trades. Test a longer history.

No edge

Doesn’t work as written

Lost money on average after costs.

Markets

US stocks, ETFs and indexes, plus crypto and forex with their own cost model. Futures and commodities are not supported yet.

Candles

5-minute, 15-minute or 1-hour rules, between 9:30 am and 4:00 pm ET.

History

3, 6, 12 or 24 months. Support and resistance rules go back 12 months.

Costs

Two cents a share round trip on stocks and ETFs, 0.1% on crypto and 0.02% on forex, taken off every trade.

Agents

Up to 5 running at once on Pro. Each needs a backtest of at least 3 months and 10 trades first.

Fair use

15 backtests an hour and 60 a day.

Questions

AI trading bot FAQ

Can I create a trading bot with AI and no coding?

Yes. In Senntric’s Strategy Lab you describe the strategy in plain English, Sena turns it into exact rules you can edit with dropdowns, you backtest it after costs, then turn it on as an agent that trades on paper, all without writing code.

Can I build a trading bot with ChatGPT or Claude?

You can ask a chatbot to write bot code, but you still need market data, a backtest engine that handles fills and costs honestly, and a way to run it live. Strategy Lab does those parts for you: the AI writes the rules, the engine tests them on real candles after costs, and an agent runs them on live prices. Senntric is also adding an MCP connector, so you will be able to check and manage your agents from Claude, ChatGPT or Cursor.

Does the trading bot use real money?

No. An agent trades a paper account of its own on live prices, sized by the account and risk you choose. Senntric never places real orders or connects to a broker.

How do I know if my trading bot works?

Look at the result after costs, the number of trades and Proven Edge. Strategy Lab needs at least 30 trades before it judges an edge, and calls a strategy “Works, with an edge” only when the conservative estimate stays above zero. Once it runs, the Agents tab shows whether its paper results stay in the backtest’s range. Forward testing vs backtesting.

How many trading bots can I run?

Up to 5 agents at once on Senntric Pro, each on its own paper account. Each one checks its rules on every candle close in regular hours.

What strategies can I build?

Strategies built from levels (opening range, support and resistance, prior day, premarket, VWAP, moving averages), entries (break, touch, bounce, reclaim), candle confirmations, up to four filters, stops and targets. Many common day-trading strategies fit.

Try it in Senntric

Test your bot before it costs you.

Strategy Lab turns one sentence into rules and a backtest after costs, then runs it as an agent on a paper account of its own.

Get startedThe full Sena guide

More Sena tutorials

Educational, not investment advice. Past results do not predict future results. See also Do AI trading bots work? and the Sena user guide.