Datapult Planning AI is a constraint-based scheduling engine. You send us employees, shifts, and rules-we send back an optimized schedule.
If you've built workforce management software, you know the pain: shift scheduling looks simple until you factor in labor laws, employee preferences, skill requirements, team constraints, and the thousand edge cases your customers throw at you. Writing a solver from scratch is a multi-year project. Maintaining it is worse.
That's what this API solves. You keep your domain logic, your UI, your customer relationships. We handle the NP-hard optimization problem.
What this API does
You POST a JSON payload describing:
- Employees with skills, contracts, and availability
- Shifts that need to be filled
- Spots (locations or roles) with their requirements
- Settings that control how we weigh competing constraints
We run a constraint solver and POST the solution back to your callback URL. The whole thing is async-submit and forget, then handle the callback when we're done.
Response times depend on problem size, but most schedules solve in under 10 minutes. Complex multi-week schedules with hundreds of employees might take longer.
Who uses this
The API is domain-agnostic. If you're scheduling people to work, it probably fits:
Retail - Store coverage across opening hours, skill-based roles (cashier, stock, supervisor), part-time vs full-time balancing.
Healthcare - Nurse rostering with strict labor law compliance, skill requirements (RN, LPN, CNA), mandated rest periods, weekend rotation fairness.
Call centers - Coverage optimization across time zones, break scheduling, skill-based routing alignment.
Hospitality - Kitchen, front-of-house, housekeeping with varying shift lengths and split-shift constraints.
Manufacturing - Line coverage, certification requirements, seniority-based assignments.
The common thread: you have people, time slots, and constraints. We handle the matching.
Integration pattern
A typical integration looks like this:
- Your system collects scheduling data (employees, shifts, preferences)
- You transform it to our request format
- POST to our API with your callback URL
- We solve and POST results back
- You apply the solution in your system
The API is stateless from your perspective. We don't store your employee data beyond what's needed to solve the current request. Each solve is independent.
Getting started
Authentication - Get your API token and learn how to authenticate requests.
Understanding the request - See the full payload structure and how the pieces fit together.
Employees - Model your workforce with skills, contracts, and preferences.
Shifts - Define the work periods you need covered.
Settings - Tune the solver's priorities for your use case.
A note on complexity
Scheduling is a hard problem in the computer science sense (NP-hard, specifically). Perfect solutions don't exist for non-trivial inputs-there are too many combinations to evaluate exhaustively.
What you get is a good solution. Often very good. The solver uses metaheuristic algorithms (constraint propagation, local search, tabu search) to explore the solution space efficiently. It won't always find the global optimum, but it'll find something reasonable in reasonable time.
If the result isn't what you expected, check the settings. Constraint weights are the primary tuning mechanism-they tell the solver what matters most to your business.
Questions?
Reach out at [email protected]. We're happy to help with integration questions, constraint modeling, or weird edge cases.