Architecture
How
works
Two layers. One API. A whole application, screens and processes, generated from metadata.
Your database (Oracle / MySQL / MariaDB)
│
▼
ApiLogicServer (ALS) — Python, open source
Reads your schema → exposes REST automatically
GET /api/customers, POST /api/orders, ...
│
▼
e-cometa — Spring Boot + Vaadin
Reads the ALS metadata and its own
Generates grids, forms, filters, menus
and runs the business processes
│
▼
The user's browser
A complete application. Nothing to install.ApiLogicServer: your database turned into a REST API
ApiLogicServer (ALS) is an open-source Python project that reads your database schema and automatically exposes a REST API with pagination, OData filters, FK relationships and column metadata.
Not a single endpoint has to be written. If your database has 50 tables, you have 50 REST resources working in minutes.
ALS auto-generates:
- GET /api/customers?filter=...
- POST /api/customers
- PUT /api/customers/123
- GET /meta/@tables/customers
- → columns, types, FKs
e-cometa generates dynamically:
- — Grid with pagination and filters
- — Edit form with validation
- — ComboBox with FK lookup
- — Navigation menu
- — Subgrids and hierarchies
- — Tabs and detail panels
- — Multi-step business processes
: the UI generated from the metadata
e-cometa reads the metadata exposed by ALS (column names, data types, relationships between tables) and automatically generates the complete user interface: grids, forms, filters and navigation.
There are no screens "saved" in the code. There is metadata that the engine interprets at runtime. That is what makes hot deployment possible, and it applies to business processes just the same.
What the application does, in metadata too
A process is a sequence of steps defined in metadata: it reads records, asks the user for whatever is missing, validates conditions, creates or modifies data, invokes a service and delivers a result, be it a report, an XML file, an email or a label.
The order comes from the sequence of the steps, and any step can jump to another depending on its outcome, so the flow branches without writing code. If a step needs a decision from the user, the process stops at a dialog and carries on where it left off.
It is triggered from a button on the grid or the form, on saving a record, when a report finishes, or from another process in a chain.
Process "Confirm order":
- 1. read · order lines
- 2. validate · enough stock
- 3. ask · delivery date
- 4. create · delivery note + lines
- 5. modify · order → confirmed
- 6. output · PDF to the customer
Six metadata records. Zero lines of code.
The life cycle of a change in production
- 1
The manager finds they need to add a "Serial number" field to the delivery notes form.
- 2
They open MetaConfig (the e-cometa metadata editor) and add the field with its properties: label, type, validation.
- 3
They save the change. The metadata is recorded in the database.
- 4
The next user who opens the delivery notes form sees the new field. Nobody else did anything.
Total time: under 2 minutes. No build. No deployment. No service interruption.
Supported databases
Oracle
Including legacy schemas decades old
MySQL / MariaDB
Recent and inherited versions
PostgreSQL
Through ALS — coming soon
Security
Spring Security — enterprise authentication and authorisation
LDAP / Active Directory — optional integration through configuration
HTTPS mandatory — no exceptions
Sensitive data masking, configurable per field
PWA — secure mobile access with no app to install
On-premise — your data never leaves your infrastructure
Want to see it running on your database?
Within 48 hours we assess your case and prepare a demo with your real data.
Request a demo →