Bring your data
A CSV from your computer, a real public sample, or a live feed. Columns get a type — time, number, text — and you can correct any guess.
T dateAa region# visits# sales
Checking the lab… · Free · no account
Research any data. No code.
Bring a file or a sample, drag steps onto a canvas — filter, group, calculate, correlate — and watch the data change at every step. Pin what you find to a DataBoard.
Try it in your browser — your file stays there unless you send it to a lane.
beesystem.io/lab
- seattle-weather.csv…
- Group & summariseby month · mean of temp_max
- Line chart…
Seattle — average daily high by month (°C)
Starting the engine in your browser…
- No account
- Your file stays in your browser
- See the data at every step
- Real public samples
What you can do
From a file to a finding, on one screen.
Five things, each one drag away.
Drag and drop
Filter, group & summarise, join, calculate, rolling windows, statistics, correlation — wired together on a canvas, no code.
Read CSVFilter rowsGroup & summariseLine chart
See every step
Click any step to see its table. Change a setting and the data follows. A wrong step says what is wrong and how to fix it.
DataBoard
Line, bar and scatter charts, tables and numbers, side by side — every one says which data and which version it came from.
Ask the assistant
Describe the question. The assistant reads your columns and proposes steps on the canvas that you keep or undo.
How it works
Three steps, no code.
What you build is a pipeline — a document of steps and wires — so it is checked before it runs, and it can be saved, shared and run again.
- 1
Bring your data
Add a file, a sample or a feed. Try it in your browser; the file stays there unless you send it to the lane.
- 2
Drag and drop
Drag steps from the library, wire them together, and click any step to see its data.
- 3
See it on the DataBoard
Pin a step and it becomes a chart, a table or a number on the board — all of them update as you edit.
Samples
Start from real data.
Public datasets with their sources. Each one opens with a working pipeline you can change.
- Seattle weatherDaily rain, temperature and wind, 2012–2015NOAA, via vega-datasets (BSD-3-Clause)Open in the workspace
- CO₂ at Mauna LoaMonthly carbon dioxide in the air since 1958Scripps CO₂ Program, via vega-datasets (BSD-3-Clause)Open in the workspace
- Global temperatureYearly land–ocean temperature anomaly, 1880–2023NASA GISS, via vega-datasets (BSD-3-Clause)Open in the workspace
- Iowa electricityElectricity generated in Iowa by source, 2001–2017U.S. EIA, via vega-datasets (BSD-3-Clause)Open in the workspace
- Health and incomeLife expectancy and income for 187 countriesGapminder (CC BY 4.0), via vega-datasetsOpen in the workspace
- Your own fileDrop a CSV on the canvas. It stays in your browser.Try it in your browser — your file stays there unless you send it to a lane.
Packs
Packs add a field on top.
A pack brings its own data source and steps to the same framework. The first one is for markets.
Market pack — the strategy lab
Public candles for six crypto pairs, 22 indicators, backtests next to buy-and-hold with fees and slippage, and pipelines that run live on every new candle.
Open the strategy labbeesystem.io/lab
BTC/USDT · 1hEMA 20 × 50Running the template on the lab…
Strategy templates
Each one opens in the strategy lab, loaded and ready to change.
- EMA 20/50 crossTwo moving averages of the close and the marks where they cross.
- RSI 14 reversionBuy when RSI falls under 30, exit when it recovers above 55.
- Bollinger bands 20/2Buy under the lower band, exit at the average.
- Donchian 20 breakoutBuy the 20-bar breakout, exit under the 10-bar low.
- MACD 12/26/9The MACD line, its signal and the histogram in their own pane.
- ATR channelA 20-bar EMA with bands two average true ranges away.
- Volume z-scoreHow unusual each bar’s volume is against the last 50 bars.
For agents
Your agent can use the lab too.
Over MCP, an agent reads the strategy lab's node catalog, validates and previews pipelines, backtests, deploys live and reads the output — the API the board uses, under the same limits.
MCP endpoint
https://lab-api.beesystem.io/api/lab/mcp- lab_catalog
- lab_validate
- lab_preview
- lab_backtest
- lab_deploy
- lab_undeploy
- lab_read
Lanes
Where your research runs.
Your own files run in your browser. Live pipelines and bigger data run on a lane — its own control plane, hive and storage.
Shared lane
Free · open now- No account needed
- Your files are processed in your browser, up to about 300,000 rows
- Up to 3 live market pipelines per visitor
- 20 market previews or backtests a minute
Dedicated lane
Coming- Your own hive and your own database
- Placed on the server you choose
- Bigger files, scheduled runs and links you can share
- Custom code nodes, in a sandbox
Your first chart takes about ten seconds.
Open a sample, click a step, change one setting. No account, nothing to install.
Start with a sample