Custom transport planning API development
Appfront builds custom transport planning APIs for logistics and transport companies that want to automate their trip planning, fleet planning and route optimisation. Under the hood: VRP solvers based on OR-Tools, proprietary ETA models, geocoding and rate shopping. On the outside: a clean REST API that connects seamlessly to your TMS, telematics, ERP and customer portals.
What is a transport planning API?
A transport planning API is a programmable layer around your planning logic. Instead of planners manually allocating orders across vehicles, or getting stuck in a closed SaaS tool, you get your own API endpoint that receives orders, vehicles, drivers and constraints and returns an optimised set of routes within seconds. Under the hood run geocoders, a Vehicle Routing Problem solver and ETA models, tailored to your type of transport and your business rules.
For many carriers this is a logical next step after choosing a TMS. A TMS organises orders, invoicing and driver dispatch, but the optimisation itself is often based on rules of thumb and manual corrections. A planning API takes over that optimisation and feeds the results back to the TMS, without you having to replace your TMS. We often see this pattern in projects around custom transport software: the TMS remains the operational core, with the planning API running alongside it as a smart companion.
Appfront designs these APIs so that they never become a black box. The solver settings, business rules, model versions and evaluation metrics can be inspected by your own planners and IT team. The integrations run via standard REST and webhooks, documented according to OpenAPI, so that external parties, such as a new customer portal or a new telematics supplier, can connect to them almost immediately. You can read more about this approach on our page about smart API integrations.
VRP solver, not manual shuffling
OR-Tools or VROOM calculates the optimisation using constraints that your planners normally keep in their heads: time windows, weight classes, driver skills, low-emission zones, ADR. The solver delivers a complete plan within seconds, and planners keep the exceptions.
ETA based on your own history
Not a generic Google ETA, but an arrival-time model trained on your own trip logs, vehicle types and routes. Waiting times at loading and unloading addresses, busy corridors and seasonal effects are built into the model, so that customer ETAs are accurate.
Open integration with your TMS
The API runs alongside Transics, Carrierweb, i-teq or another TMS. Orders and vehicles are read in via REST or message queues, and the optimised plan flows back. Your TMS remains the operational source, while the optimisation becomes visible and controllable.
Our development process for a planning API
A good planning API doesn't start with code, but with the planners who currently do the work. Together with them, we bring the real constraints to the surface, often unwritten knowledge that has sat in the heads of senior planners for years. Only then do we choose the solver, the data model and the integrations. That way you get an API that is not only mathematically optimal, but also earns the trust of the team that has to use it every day.
We spend several days alongside your planners, mapping out constraints, exceptions and KPIs. Without that context, any VRP solver fails in week one.
OR-Tools, VROOM or a hybrid approach, depending on scale, constraints and the computation times you need. We define the data model and the OpenAPI contracts.
Iterative development with automated tests and shadow runs alongside your current planning. Planners see results and give feedback before we go live.
Controlled rollout per depot or region, with monitoring of solver quality, ETA errors and exception rates. After that comes ongoing maintenance and refinement.
What a planning API delivers in practice
Every planning API is tailored to the type of transport and the existing systems of the carrier. Below are the components we consistently see, which can be combined as needed depending on your scenario. What this page covers ties in with the broader context we describe in what is an API integration, translated to the specific challenges of transport.
Route planning & VRP solver
The solver allocates orders, vehicles, drivers and restrictions across routes at the lowest cost or CO₂ emissions. Time windows, capacity, skills, ADR and low-emission zones are built into the model as constraints — your planners handle the exceptions, while the solver does most of the work.
Geocoding & address validation
Order addresses are converted to coordinates via PDOK, Photon or Nominatim, corrected for spelling errors and validated against postcodes and BAG registers (the Dutch address and building registry). From a single API endpoint, you get reliable lat/lon pairs the solver can work with.
ETA prediction
An arrival-time model trained on your own historical trip data, supplemented with traffic sources and weather information. Waiting times at specific loading and unloading addresses, weekend effects and seasonal patterns are built into the model, so customer ETAs are accurate and planners handle fewer calls.
Rate shopping & carrier selection
For shippers and logistics service providers that use multiple carriers: rates from charter, scheduled-service and parcel carriers are retrieved, validated and compared in real time. The API selects the best combination of price, transit time and reliability based on your own criteria.
Capacity forecasting
Based on historical order patterns and seasonal effects, the API predicts how many vehicles and drivers you will need in the coming days or weeks. Staffing and charter planning thus becomes data-driven rather than reactive — especially valuable for seasonally sensitive sectors.
Exception handling
Traffic jams, no-show customers, vehicle breakdowns, last-minute orders — in transport, exceptions are the rule. The API offers re-planning endpoints that can quickly generate an adjusted schedule without replanning the entire day. Planners stay in control, while AI does the heavy lifting.
Typical scenarios in practice
A transport planning API looks different for each type of business. We see a number of scenarios recur structurally — for each of them, we have a recognisable approach that takes into account the specific rules, restrictions and systems common in that market. Also read our broader view of this sector in transport & logistics.
Distribution & FMCG
Distributors who plan hundreds to thousands of stops every day from one or more DCs. The API factors in time windows, weight, volume and vehicle type. Results flow back into Transics or a comparable TMS, so planners can focus on exceptions instead of shifting jobs around manually.
Last-mile & e-commerce delivery
Online retailers' delivery operations and last-mile carriers that promise consumers time slots and have to deliver on them. The API combines VRP with real-time ETA prediction and sends slot suggestions back to the customer portal, so the slots on offer and operational reality stay in sync.
Food & pharma with cold chain
Temperature-sensitive logistics where cooling requirements, four-eyes checks and HACCP or GDP requirements are part of the planning. The API knows vehicle cooling compartments, rules on what may not be combined, and specific loading and unloading sequences — so compliance is built into the route choice from the start.
Construction logistics & project transport
Construction logistics with arrival windows at the building site, permit documents, low-emission zones and ADR requirements. The API is based on resource-driven planning, where each trip often has its own story — suited to freight forwarders who do more than standard scheduled transport.
Technology we use
We tailor the technical stack to your scale, constraints and existing IT landscape. We use proven open-source components where possible and our own models where they add value. The API itself is built in a language and framework that your own team can continue to maintain, with no vendor lock-in and no secret sauce.
If you want to see how full each trip, customer and vehicle runs, read more about software for trip planning and load factor.
Why Appfront for your planning API?
Appfront has extensive experience building APIs for transport and logistics, as well as with the mathematics behind planning. We know that a VRP solver only becomes useful once planners trust it, that an ETA model only works if it is trained on your own data, and that an API is only valuable if it fits in with what is already in place. That is why we always start with your planners and your existing TMS, not with a blank solver sheet.
For every integration, we write clear documentation and OpenAPI contracts so that your own team or a future supplier can understand and manage what is running. No black box, but transparent code and clear agreements on solver settings, model versions, monitoring and maintenance. That way you stay independent and can keep control of your planning logic at any time, a requirement that is also becoming legally more important under the AI Act for algorithms that direct people or vehicles.
You work with a single point of contact who understands both the optimisation and the integration side. This keeps lines short, prevents miscommunication between your TMS supplier, telematics partner and in-house IT, and speeds up decisions when choices have to be made during the build about data models, constraints or ETA models.
Also take a look at our broader overview of the integrations we build and how we approach API integrations in general. The planning API is a specialised application of that.
- Experience with OR-Tools and VROOM in production
- Specialist in transport and logistics software
- Custom ETA models based on your historical data
- Familiar with Transics, Carrierweb and i-teq adapters
- Telematics integrations with Webfleet, Trimble, Geotab, MiX
- Geocoding based on PDOK, Photon and custom corrections
- OpenAPI 3 contracts, no undocumented endpoints
- Secure by default: API keys, scoped permissions, rate limiting
- EU-resident hosting (Azure West Europe, AWS Frankfurt, Dutch cloud)
- AI Act preparation for scheduling algorithms
- A single point of contact, no passing you between account managers
- Ongoing maintenance, monitoring and model renewal
Security, GDPR, EU residency and the AI Act
Transport data is sensitive: driver personal data, GPS tracks, driving and rest times and order information fall under the GDPR as well as sector-specific rules. The algorithms that act on this data also fall under the European AI Act. Appfront builds in line with the OWASP ASVS and, for every planning API, explicitly determines which personal data is needed, who has access and how long it is retained.
By default, we host within the EU (Azure West Europe, AWS Frankfurt or a Dutch cloud of your choice) and encrypt data in transit and at rest. We document the data flows between your TMS, telematics, customer portal and planning API so that your record of processing activities is complete and you can demonstrably comply with the GDPR. For driving and rest times, we take EU Regulation 561/2006 into account: data is not retained longer than legally permitted and is only accessible to roles that genuinely need it.
For the AI components (VRP solver, ETA model and any assignment algorithms), we provide model cards, a description of where the training data comes from and evaluation metrics. This allows you to demonstrate under the AI Act that decisions are transparent and explainable. For algorithms that direct drivers, such as workload balancing, we also provide an impact analysis and human oversight steps, in line with the requirements for high-risk AI. Also read our information security policy and coordinated vulnerability disclosure policy.
- GDPR-compliant processing of driver and customer data
- EU residency: Azure West EU, AWS Frankfurt or a Dutch cloud
- AI Act: model cards, dataset provenance, evaluation metrics
- Impact assessment for algorithms that direct drivers
- EU 561/2006: driving times retained for the statutory retention period
- eCMR-compliant digital consignment notes with audit trail
- Encryption in transit (TLS 1.2+) and at rest
- API key rotation and role-based access
- Rate limiting and DDoS protection
- Audit logs with traceable data flows
- Monitoring and alerting for solver and model drift
- Documentation for your record of processing activities
Frequently asked questions about a transport planning API
Answers to the questions we are asked most often about custom planning APIs for transport and logistics.
A transport planning API is a custom, programmable layer around your planning logic. Instead of manually dragging trips around in a TMS screen or planning via a closed SaaS button, the API offers endpoints for entering orders, vehicles, drivers and constraints, and returns optimised routes as a structured response. Under the hood run geocoders, a VRP solver (for example OR-Tools) and ETA models that take into account driving times, time windows, capacity and historical traffic patterns.
SaaS tools such as Routigo, PTV Smartour or Stratumn work well for standard distribution, but often hit their limits on specific constraints: weight classes, low-emission zones, permit documents, refrigerated routes, four-eyes checks, ADR, partial loads or customer-specific loading arrangements. Your own planning API gives you control over the optimisation goal, the business rules and the way the solution connects to your own TMS, ERP and driver apps. In addition, the data stays within your own domain, which is increasingly important under the AI Act for algorithms that steer people or vehicles.
For the VRP part we typically work with Google OR-Tools (CP-SAT and the Routing Library) or with VROOM, depending on the scale and the type of constraints. For pure shortest-path calculations and isochrones we use OSRM or GraphHopper on top of a current OpenStreetMap extract; for geocoding we combine local geocoders with Photon/Nominatim or the PDOK location server for Dutch addresses. We train ETA prediction as a gradient-boosting model or a lightweight neural network on your own historical trip data, supplemented with external traffic and weather series. Which combination fits best is something we determine after a short data discovery.
We build the API as an independent service that talks to your existing systems via REST, webhooks and, where necessary, message queues. For Transics, Carrierweb, i-teq or any other TMS, we provide adapters for orders, vehicles, drivers and planning results. For telematics (Webfleet, Trimble, Geotab, MiX) we ingest position and driving-behaviour data in real time. For ERP and accounting we connect using the standard connectors (SAP, Microsoft Dynamics, Exact, AFAS). This keeps the planning API as the only place where the optimisation logic lives, leaving the rest of your IT landscape untouched.
The cost is determined by the complexity of your constraint set, the number of orders and vehicles you plan each day, the computation times you need, the number of systems to integrate and the extent to which you want your own models for ETA, demand forecasting or carbon reporting. Monitoring, maintenance and the degree of self-management also play a role. After a no-obligation analysis, we put together a clear quote covering the scope and any phasing.
Yes. We host the API within the EU (Azure West Europe, AWS Frankfurt or a Dutch cloud of your choice) and build in line with OWASP ASVS. Driver data, trip logs and GPS tracks fall under the GDPR and are retained to a minimum, encrypted in transit and at rest. For the AI components (the VRP solver, the ETA model and any allocation algorithms) we document model cards, training data provenance and evaluation metrics, so that under the AI Act you can demonstrate that decisions are transparent and explainable. For algorithms that direct drivers, we also provide an impact assessment and human oversight steps.
Yes. Appfront regularly takes over the management of existing planning services, some originally written by in-house IT and some by another supplier. We review the solver configuration, business rules, API contracts and monitoring, document the current setup and draw up an improvement plan. From that point on, we can take on extensions, performance tuning and ongoing maintenance, including rolling out new models for ETA or capacity forecasting.
A transport planning API pays off most when you plan large numbers of orders, vehicles or stops every day and standard SaaS tools no longer cut it. Typical profiles: mid-sized to large road hauliers, distribution and delivery companies, couriers with their own consumer portal, construction logistics, food and pharma logistics with strict time windows, last-mile delivery companies and logistics service providers with multiple clients per trip. For smaller fleets, a focused, lighter VRP service may be enough.
Talk to us about your planning API
Tell us how you plan now, which TMS and telematics you work with and what you would like to automate: VRP, ETA, rate-shopping, capacity forecasting or exception handling. We are happy to help think through scope, priorities, data model and alignment with your existing suppliers. A no-obligation first conversation gives you a clear picture of what is feasible for your situation.