How to build a mobile app for TOTVS Protheus
Short answer
Build a mobile or rugged-handheld app that talks to Protheus through REST endpoints and ExecAuto routines for receiving, stock moves, production reporting and approvals. Plan for barcode scanning, intermittent connectivity and a VPN or gateway to an on-premises AppServer.
Protheus screens are built for the desktop. On the warehouse floor people end up writing numbers on paper and typing them in later, which is where stock inaccuracy starts. A small app for the few moves that matter can fix the largest share of that.
TOTVS has offered mobile and warehouse-oriented options over the years, and some installations use them. Many plants still need something tailored to their layout, labels and routines, built on the same REST and ExecAuto layer.
The principle is the same as in any ERP mobile project: pick three or four transactions, make each take under thirty seconds, and keep all business rules in Protheus so the app stays thin.
What teams use it for
Receiving against purchase orders
Scan the supplier label or order number, confirm quantities per line and send to a pre-invoice or receipt, with differences visible at the dock.
Warehouse transfers and picking
Scan origin, item and destination to post internal transfers, and follow pick lists for sales orders with scan-to-confirm to cut wrong picks.
Production order reporting
Operators scan the order and report quantity produced, scrap and time per operation, feeding production appointments (apontamentos) and the costing downstream.
Physical inventory counting
Counters scan location and item and enter quantities against the inventory count in Protheus, with recounts triggered when variance exceeds a limit.
Approvals on the move
Managers approve purchase requests and orders from their phone with the same approval authority they have on the desktop.
TOTVS Protheus integration surface
- Protheus REST services on the AppServer, standard TOTVS APIs and custom TL++ endpoints, as the main contract for the app, with JSON payloads.
- ExecAuto routines for stock movements (internal movements and transfers), receipts and production appointments, so rules and costing run as on the desktop.
- Authentication options on the AppServer REST layer (basic or token-based, depending on release), plus TOTVS Fluig or an identity provider in front if you need SSO.
- Product barcodes and complementary tables (such as product by supplier and address/location structures under warehouse management), used to resolve scans to items and locations.
- Direct read-only database views for fast lookups of stock by location when the REST endpoint would be too slow for scanning.
How it works
- 1
Choose devices and platform
Rugged Android handhelds with built-in scanners suit warehouses. Phones work for approvals and light tasks. A cross-platform framework avoids two codebases.
- 2
Define the API contract per transaction
Write one endpoint per task with the minimum fields, and return clear error codes that the app can show in plain language to the operator.
- 3
Resolve scans on the server
Send the raw barcode and let the server resolve item, lot, location or order, so label formats can change without an app release.
- 4
Design for weak connectivity
Queue transactions locally with a unique id, sync when online, and make the endpoints idempotent so a retry never posts twice.
- 5
Reach the AppServer safely
Use the plant wifi to the internal network, and a VPN or mobile device management profile for remote users. Do not publish the AppServer directly.
- 6
Pilot on one area
Start with receiving or one production line, measure time per transaction and error rate, then widen.
Example flows
Dock receiving with a handheld
Trigger: A truck is at the dock with goods for an open purchase order.
- Operator scans the order barcode; the app loads open lines from Protheus.
- Each item is scanned and a quantity is entered or counted.
- Differences from the order are highlighted before submission.
- The app posts a pre-invoice draft through the server routine.
Result: Receiving is recorded at the dock, with variances caught before the truck leaves.
Production reporting from the line
Trigger: An operator finishes a batch on a production order.
- Operator scans the order card and selects the operation.
- Good quantity, scrap and reason code are entered.
- App queues the report if the network is weak and syncs when it returns.
- Server posts the appointment through the production routine and returns the new balance.
Result: Production status is current within minutes instead of at the end of the shift.
Keeping data on-prem
Most plants running Protheus on-premises are comfortable keeping the mobile layer there too. Host the API gateway inside the network, let handhelds join the plant wifi, and reach the AppServer over internal addresses only. For off-site users, a VPN or an MDM-managed tunnel is safer than a public endpoint. If you add AI on top, such as voice entry, a photo-based count or a natural-language assistant for supervisors, run the models on local GPUs and have the app call a service in your network. Nothing about a barcode transaction requires the cloud, which also means the app can keep working through an internet outage as long as the local network and AppServer are up.
Security checklist
- Authenticate each user with their own credentials or SSO token, not a shared device login, so postings are attributable.
- Enrol devices in MDM with a screen lock, remote wipe and a managed app configuration.
- Use TLS on the internal API as well, and pin certificates in the app where practical.
- Limit each role to its transactions on the server, and log device id with every posting.
Ways to get it done
TOTVS mobile or warehouse offerings, if licensed
Effort: Configuration and training, with limited customisation.
Standard receiving, picking and counting when the vendor layouts fit your process.
Custom app on REST and ExecAuto built in-house
Effort: Two to four months for a first set of transactions.
Teams with a mobile developer and TL++ capacity and layouts that differ from standard.
Partner-delivered app plus AI assist, such as Netray
Effort: Pilot in one area, then scale by transaction.
When you want voice, photo counting or a supervisor assistant over live data with models kept private.
Pitfalls
- !Copying the desktop screen flow instead of designing a task that fits a scan and two taps.
- !Posting non-idempotent transactions so a retry after a timeout doubles a stock movement.
- !Resolving barcodes in the app, which makes every label change a release.
- !Ignoring lot, serial and location control that the desktop enforces and the endpoint must too.
Frequently asked questions
Does TOTVS Protheus have a mobile app?
TOTVS has offered mobile and warehouse-oriented products for Protheus across releases, and availability depends on your version and licences. Many plants add a custom app on Protheus REST for their own transactions. Check with TOTVS or your partner before choosing.
How does a mobile app post transactions to Protheus?
Through REST endpoints on the AppServer that call ExecAuto routines, so the normal validations, costing and fiscal rules apply. Avoid direct database writes, which skip those rules and are hard to support.
Can the app work offline?
Yes, with a local queue and idempotent endpoints. Reads such as order lists can be cached, while writes sync later. Decide which transactions must be confirmed live, such as those that reserve stock.
How do handhelds reach an on-premises AppServer?
Over the plant wifi to an internal gateway, and through a VPN or MDM tunnel for remote users. The AppServer itself should not be reachable from the public internet.
Is an AI assistant useful on the shop floor?
Sometimes. Voice entry and photo counting can save time, and a supervisor assistant can answer shortage questions from live data. Keep it read-only first and run the model on-prem where plant data is sensitive.
Want Mobile app working on your TOTVS Protheus?
A free 30 minute assessment: we map the integration path for your version and setup, and where it should run.
Get a free assessment