A self-hosted WMS runs in an environment your business or appointed partner operates, on local servers or in a cloud account. It can give you more control over deployment, access and change timing. It also leaves operating work that someone must own and fund.
The first question is practical: if picking stops tomorrow, who investigates the application, restores the database and gets the floor working again? A low server bill does not answer that question.
This guide separates deployment, software rights and daily responsibility so you can compare a self-hosted warehouse management system with a managed product. Product documentation was checked October 22, 2026. The suggested tests are planning examples, not reported warehouse results.
Separate hosting, licensing and ownership
These are three decisions, even when one supplier provides everything.
| Decision | What to establish before buying |
|---|---|
| Where it runs | Your premises, a customer-controlled cloud account or a provider-managed environment. Confirm the supported architecture. |
| What you may do | The licence permits your deployment, users, locations, modifications and continuing use under its actual terms. |
| What you can access | Application code where agreed, database records, configurations, documents, backups and the tools needed to use them. |
Self-hosting does not require owning the source code. Proprietary software may be licensed for customer-operated deployment. Open-source software grants rights under its licence; it does not transfer the original project's copyright to you. A custom contract can assign rights to the commissioned work while third-party components retain their own licences.
For a concrete commercial example, Odoo's Enterprise agreement permits self-hosting under a licence to use the software for the agreement's duration. Hosting it yourself is therefore different from owning it outright. Check the relevant product and edition rather than assuming that on-premise means a perpetual or unlimited-user licence.
Data access is a separate issue too. Confirm export coverage and continuing access in either model. A SaaS contract can provide data-export rights; a self-hosted database still needs documentation to preserve usable client, stock and order relationships.
Which self-hosting route are you considering?
There are three reasonable starting points:
- A licensed product with a supported deployment option. Evaluate warehouse fit, licence limits, upgrades and the support arrangement together. You may still pay recurring software and maintenance charges.
- An open-source application. Check receiving, allocation, scanning, shipping and reporting against your actual work. Then price configuration, integrations, upgrades and support. Source availability is useful, but does not complete implementation.
- A custom application. Define the workflow, ownership and operating plan before commissioning it. A focused portal or integration is a different undertaking from replacing warehouse execution.
OpenBoxes' pricing page illustrates the distinction between free community software, self-hosting and managed hosting. It explicitly notes operating costs and separately describes implementation services. Its installation documentation covers configuration, databases, deployment and upgrades. Those are real responsibilities behind the download button.
A home-inventory application is not automatically suitable for a commercial warehouse. Demonstrate client separation, concurrent work, stock corrections and the shipping process you need. If you operate a 3PL, also check client-level visibility, billable activity and the records finance requires.
For a product evaluation around those jobs, use our ShipHero and Extensiv alternatives guide. It compares warehouse tasks and quote questions rather than treating every inventory application as an equivalent WMS.
Choose the environment around the operating requirement
| Environment | A reason to investigate it | Work to include in the plan |
|---|---|---|
| Customer-controlled cloud account | You want account access, a specified deployment region and a clear route to another operating partner. | Service configuration, permissions, database and application care, monitoring, backups and usage charges. |
| Servers at your facility | A supported application and local infrastructure fit the operation's connectivity or deployment requirements. | Hardware lifecycle, power, networking, physical access, off-site recovery copies and staff or partner coverage. |
| Partner-operated deployment | You want agreed software and infrastructure work handled by an operating partner. | Account access, included tasks, support hours, recovery arrangements, exclusions and a usable handover. |
A cloud account in your company's name still uses a third party's infrastructure. Carrier, commerce, accounting, email and monitoring connections may also move data outside that account. Map those flows rather than promising that no third-party server is involved.
On-premise hosting also does not automatically make a system work offline. Test what the chosen application and connected services can do during an internet outage, including shipping labels and later reconciliation. Local servers can remove one dependency while leaving others.
Agree who owns each operating task
AWS's shared-responsibility guidance explains that responsibilities vary with the services selected. A managed database can reduce some infrastructure work, but it does not remove responsibility for application behaviour, permissions or data. An application running on a virtual server can leave more operating-system work with the customer or appointed partner.
Use the proposal to assign these tasks, not just label a line “hosting and support.”
| Task | A useful acceptance question |
|---|---|
| Accounts and permissions | Who approves access, removes departing staff and holds recovery access to the hosting account? |
| Updates and dependencies | Who tracks supported versions, tests a change and schedules its release around warehouse work? |
| Backups and recovery | Who checks that copies can be restored, which records are covered and how much recent work might need reconciliation? |
| Monitoring and incidents | Who receives an alert, during which hours, and who investigates application, infrastructure and integration faults? |
| Connected systems | Who resolves a failed or duplicated order update and checks the resulting stock and shipment state? |
| Handover and continuity | Could another qualified operator obtain the deployment instructions, configuration and permitted access? |
Specify response arrangements separately from recovery expectations. A promise to acknowledge an incident is not a promise to restore shipping within the same period. Internal staff still need to approve changes, manage users and reconcile exceptions even with a support partner.
Security and compliance depend on the implemented controls, actual data and applicable requirements. A selected region or hosting provider's certification does not certify the whole WMS. Establish the required access, retention, logging, recovery and supplier controls for your operation, then check the architecture against them.
Cost the operation, not just the server
There is no defensible universal monthly total from the words “self-hosted WMS.” A small test instance and a production warehouse with recovery capacity are different scopes.
Ask for a budget with these lines visible:
| Cost group | What belongs in it |
|---|---|
| Starting work | Licence or development, configuration, integrations, data preparation, migration, testing and training. |
| Infrastructure | Application and database resources, storage, network traffic, backup retention and any recovery environment. Local hardware also needs replacement planning. |
| Software and external services | Recurring licences, carrier or commerce services, integration tools and retained subscriptions. |
| Technical operation | Application maintenance, monitoring, security updates, backup checks, recovery exercises and incident coverage. |
| Internal work and change | Staff administration, exception handling, approval time, refresher training and future workflow changes. |
For the first year, add starting work to the expected twelve months of operating costs and internal time. For later years, remove genuinely completed setup costs, retain support and licences, and add planned upgrades or replacement work. Use a cloud estimate for the proposed architecture and normal and peak workloads, not a server price copied from an unrelated deployment.
Count every item once. If a support quote includes monitoring or backups, do not add the same service again. If the existing WMS remains, keep its subscription. Staff doing the maintenance themselves have a time cost even when no supplier sends an invoice.
Self-hosting also does not establish no per-user charges. The software licence, connected tools or support agreement may still use that model. Our flat-fee WMS pricing guide addresses account prices, seasonal users and contract limits separately from deployment.
Test one warehouse cycle and one recovery
Before choosing an operating model, ask the proposed product or development partner to demonstrate a representative receipt, putaway, allocation, pick, shipment and return. Include the scanners, labels and integrations the floor actually uses. Have an operator and supervisor test a correction, not only the ideal sequence.
Then test a recovery in a separated environment. Restore copied records, check stock and an open order, and verify the configuration needed for the next step. Prevent that test environment from sending real orders, labels or customer notifications. Agree the recovery target and the checks needed before returning to live work.
For a migration, identify the export/import coverage for clients, SKUs, locations, stock, open orders, returns and relevant history. Reconcile a pilot and document the cutover decision. One system must control live stock and shipping at each stage; a parallel comparison should use copied records or a clearly separated pilot.
These demonstrations expose different risks. A product may cover the warehouse workflow but need a stronger operating partner. A well-managed server may host software that cannot perform the required job. Neither test substitutes for the other.
When self-hosting is a good next investigation
Investigate it when you can name a deployment, integration or change-control requirement that the current arrangement cannot meet, and you can fund the people and systems needed to operate it. Confirm the supported product and licence before assuming a custom build is necessary.
A managed SaaS product can be the better fit when it handles the warehouse work and the provider's documented access, export and support arrangements meet your requirements. There is no user-count threshold that settles this decision for every warehouse.
If only client reporting, billing checks or an integration needs work, consider improving that part while retaining reliable warehouse execution. Ekyon's custom 3PL software service covers focused portal, billing and integration scopes. Suitable focused projects start at US$5,000; a full WMS replacement requires separate discovery and a quote. Hosting, licences, support and retained systems remain in the budget.
Bring the current platform, the requirement it cannot meet, connected systems, normal and peak work, and the person responsible for operation. That is enough to start a useful scope discussion without sharing credentials or private shipment records in an initial enquiry.
Which warehouse workflow needs a different setup?
Share the current system, the deployment constraint and the work your team wants to improve. We can discuss a focused scope, operating responsibilities and whether custom work fits.
Frequently Asked Questions
No. A commercial licence can permit customer-operated deployment without transferring copyright. Open-source and custom arrangements have their own rights and obligations. Check deployment rights, user and location limits, modifications, continuing use and access to the material needed for support.
Yes, if the chosen software supports that deployment. Your business can control an account while a partner operates agreed services. The cloud provider is still a third party, and connected carriers, commerce systems and other services may process data elsewhere.
The total includes starting licence or development work, configuration, migration, training, infrastructure, recurring software charges, maintenance and internal administration. A server bill alone is not the operating total. Obtain a quote and infrastructure estimate for the actual workflow, workload and recovery requirements.
Neither follows from the deployment label. Security depends on access, configuration, updates, data flows, monitoring and recovery. Check the implemented controls against the actual requirements. A managed SaaS arrangement can also be suitable when its documented controls and terms meet those needs.
Potentially. A portal, report or integration may be a focused project if access, records and operating responsibilities can be agreed. Keep the existing WMS subscription and any connected-service charges in the cost comparison. A full replacement is a separate scope.
Planning a warehouse software move?
See how Ekyon approaches custom 3PL platforms, WMS ownership, migration, billing, portals, and cost control.
