Usage limits
Review the limits that shape MAG OptiAI free trials, approved product beta access, AI credits, saved work, compute usage, and fair-use controls.
Plan summary
The values below show standard Free Trial limits and the Pro-level limits used for manually approved, product-specific beta evaluations. Enterprise arrangements may use separately reviewed terms.
Product limits
Inputs per run
Each workspace has its own input size, row, document, upload, or routing limits.
Saved work
Projects and records
Project, scenario, document, model, storage, and conversation limits are tracked separately from product run size.
AI credits
Separate balance
AI credits are listed separately from product, saved-work, runtime, and fair-use limits.
Product limits
These limits describe the size of a single product workflow, such as jobs, employees, stops, rows, files, pages, or upload size. Saved projects, scenarios, documents, and AI credits are listed separately below.
Production Scheduling
SchedAI
Free Trial
Pro launch
Employee Scheduling
SchedAI
Free Trial
Pro launch
Service Scheduling
SchedAI
Free Trial
Pro launch
Academic Scheduling
SchedAI
Free Trial
Pro launch
CVRP Routing
LogiAI
Free Trial
Pro launch
Forecasting
Operational AI
Free Trial
Pro launch
Anomaly Detection
Operational AI
Free Trial
Pro launch
Fraud / Risk Scoring
Operational AI
Free Trial
Pro launch
ExtractAI
Operational AI
Free Trial
Pro launch
DocAI / Knowledge Assistant
Operational AI
Free Trial
Pro launch
AI credits
AI credits are a separate allowance for supported AI actions. The free trial starts with 30 one-time ai credits, and an approved beta receives the listed allowance only for the selected product and 14-day access window. Self-service top-ups are currently paused.
Production Scheduling
300 / month
Employee Scheduling
300 / month
Service Scheduling
300 / month
Academic Scheduling
300 / month
CVRP Routing
300 / month
Forecasting
300 / month
Fraud / Risk Scoring
300 / month
Anomaly Detection
300 / month
DocAI / Knowledge Assistant
600 / month
ExtractAI
900 / month
Agentic Operations private beta
Founder and Test access is available during private beta. Free Trial and Pro limits are prepared for a later rollout but access remains blocked until the Copilot is ready for promotion.
Founder shadow training (available)
Unlimited sessions and messages; 50 files per session; 25 MB per file; 500 MB per session technical ceiling.
Test beta (available)
20 active / 100 retained sessions; 100 user messages per session; 20 files per session; 25 MB per file; 250 MB per session.
Free Trial (planned, access blocked)
3 active / 10 retained sessions; 100 user messages per session; 5 files per session; 5 MB per file; 25 MB per session.
Pro (planned, access blocked)
20 active / 100 retained sessions; 100 user messages per session; 20 files per session; 25 MB per file; 250 MB per session.
Copilot AI requests
Each successful AI-bearing request uses 1 AI credit from the account's applicable credit pool. Deterministic downloads, approvals, status checks, and result loading use 0.
Approved optimization
Creating an approved schedule uses 1 compute run, up to the approved time limit, and 0 additional AI credits.
Saved work
These limits cover what can be saved in a workspace, separate from the size of a single optimization, training, extraction, or DocAI action.
Optimization saved work
Free Trial
Pro launch
Forecasting / Fraud / Anomaly saved work
Free Trial
Pro launch
ExtractAI saved work
Free Trial
Pro launch
DocAI saved work
Free Trial
Pro launch
Retention
Retention timing is separate from saved-work capacity. Cleanup language remains cautious because records may be archived, marked for deletion, or queued for cleanup.
Free Trial
Active unused saved work may be archived after 30 days; archived saved work may be deleted after 7 days.
Test
Active unused saved work may be archived after 90 days; archived saved work may be deleted after 30 days.
Pro
Active unused saved work may be archived after 90 days; archived saved work may be deleted after 30 days.
Founder
Founder saved work is exempt from automatic cleanup.
Current/protected records
Current, staged, active, or protected state may be excluded from cleanup while needed.
Compute
These windows help set expectations for how much compute activity a workspace can run during launch.
| Plan | Active runs | Max runtime | Daily window | Weekly window | 30-day window |
|---|---|---|---|---|---|
| Free Trial | 1 | 120 seconds | 30 runs / 20 minutes | 50 runs / 45 minutes | 150 runs / 2 hours |
| Test | 2 | 300 seconds | 25 runs / 30 minutes | 125 runs / 2 hours | 300 runs / 6 hours |
| Pro | 3 | 300 seconds | 75 runs / 90 minutes | 350 runs / 6 hours | 1,000 runs / 20 hours |
Scheduling workspaces
Most scheduling workspaces expose a 300-second maximum runtime control.
CVRP
CVRP exposes a 180-second runtime cap in the Pro Console control surface.
Fair use
These caps help keep launch capacity available across customer workspaces.
| Action type | Per minute | Per day | Monthly review | Concurrent runs | File size | Pages |
|---|---|---|---|---|---|---|
| Core optimization | 5 | 100 | 1,000 | - | - | - |
| ML training | 5 | 100 | 1,000 | - | - | - |
| ML prediction/scoring | 15 | 500 | 5,000 | - | - | - |
| AI assist | 20 | 300 | 3,000 | - | - | - |
| ExtractAI | 5 | 100 | 1,000 | 2 | 10 MB | 25 |
| DocAI questions | 15 | 300 | 2,000 | 2 | - | - |
Some actions may pause when minute, daily, concurrency, file-size, or page-count limits are reached. Monthly review thresholds are used to identify unusually heavy usage during launch.
Need help choosing?
Public limits help teams evaluate fit. Billing and workspace cards in the authenticated Pro Console show the account's actual usage, remaining credits, beta requests, and product-specific access.