Quant Cloud

Quant Cloud runs forward and bulk forward algos on a remote runner so your strategies can keep executing when your PC is off or QuantCraft is closed. The desktop app connects to runners provisioned for your account after you sign in — it does not use your local Python service for these jobs.

This page covers what you do in the app. Server setup and operations are outside the scope of this documentation.

What Quant Cloud is for

Use case	Description
Bulk forward	Run a compiled algo (`.qcraft.json`) against many symbols on a schedule — for example scan a watchlist every few minutes.
Keep running when offline	Jobs continue on the remote runner after you close QuantCraft or shut down your computer.

Quant Cloud is a desktop feature. You need an active QuantCraft subscription that includes Quant Cloud slots (see your plan).

Opening Quant Cloud

Sign in to QuantCraft.
Open Auto Trading or the Quant Cloud area from the main app (exact label may vary by build).
You see runner cards for endpoints assigned to your account.

Each card represents one remote runner you can send bulk forward jobs to.

Starting a bulk forward run

On a runner card, start a new bulk forward job (or edit an existing one).
Choose your compiled algo (.qcraft.json), target symbols or scope, period/timeframe, and optional script parameters.
Optionally set timerIntervalMs if your strategy defines on_timer (Strategy lifecycle).
Submit the run.

You can Load settings… / Save settings… at the top of the modal to reuse a run preset (.qcs).

When editing an existing bulk run, the algorithm (algoId) is fixed — load settings updates targets, timer, and params but does not swap the algo file.

Add dependencies

Forward and bulk algos run in the runner's Python environment, not your local IDE .venv. If your algo imports packages beyond the defaults (for example pandas, numpy), use Add dependencies on the runner card:

Enter comma-separated PyPI package names.
The runner installs them in the background.
After a successful install, the runner restarts automatically.

While an install is in progress, you may see status in the dependencies panel. You cannot start certain operations if another install is running or a bulk job is active — wait for the current task to finish.

Install the same libraries you use in the IDE under Packages (Dependencies) when those packages are required at runtime.

Logs

Each bulk forward run streams stdout, stderr, and status events to the runner's Logs tab in Quant Cloud:

Action	What it does
Live stream	New log lines appear as the run executes.
Pause / Resume	Client-only: pauses auto-scroll and live merge; the runner keeps writing.
Refresh	Reloads the recent log tail from the runner.
Clear logs	Clears the on-screen list only — does not delete server history.

Older lines may be dropped when log volume is high. Use Refresh if you suspect you are only seeing a tail of the full history.

Deleting a bulk run removes that job's stored log file on the runner.

Script parameters and presets

Bulk forward modals use the same params.NAME discovery as the backtest dialog. Declare parameters in your strategy source, compile to .qcraft.json, then set values in the bulk forward modal.

Save complete modal setups (symbols, timeframe, inputParams, timerIntervalMs, includeFactors, etc.) as .qcs files — see Run presets (.qcs). Chart forward and bulk forward runs can use saved includeFactors / factors settings when configured in algo parameters — see Fama-French factors.

Tips

Match IDE and runner packages — if Test passes locally but the cloud run fails on ImportError, add the missing package via Add dependencies.
on_timer is for forward runs — the IDE backtest dialog does not configure timers; set timerIntervalMs in the bulk forward or chart forward modal instead.
on_tick on forward runs — fires on every broker price update from Alpaca while the algo runs; see Strategy lifecycle.
Alpaca credentials — bulk jobs that trade live or paper may require broker credentials submitted per run; follow the modal prompts for your account type.
Workspace folder — saving .qcs presets requires a linked workspace folder.

Documentation