proALPHA

• “Which data areas does our proALPHA expose through the interface?”
• “Show me the first 20 records from the Articles endpoint.”
• “What do the field names in this endpoint mean?”
• “Read out the full record with this ID.”
• “Summarize which fields the Customers endpoint returns.”

The assistant always takes the same path: first list the areas, then the endpoints in them, then their version, then the field descriptions, and only then the records. That sounds cumbersome, but it is the only honest path, because every proALPHA installation looks different.

Because proALPHA does not ship a fixed list of interfaces. The Read API is a toolkit: which endpoints exist is decided by your installation. A consultant, a partner or your own IT creates them under Data exchange, Configure REST endpoints. Two proALPHA customers therefore have two different interfaces.

What is the same everywhere is the build of the Read API itself, and luckily it describes itself. That is why the assistant first asks your system what it even offers, instead of guessing areas that may not exist for you at all.

1. 01In proALPHA open user administration and create a dedicated M2M user. A normal ERP login is not enough.
2. 02On the Permissions tab tick only REST-SERVICES.
3. 03Set a strong password and note username and password in a safe place.
4. 04Make sure that under Data exchange, REST endpoint configuration at least one read endpoint exists.

1. 01In anymize open Apps, then Connectors, and choose proALPHA.
2. 02Enter your installation address, for example https://erp.your-company.com, plus the M2M user and password from the section above.
3. 03Set which areas the assistant may use: the endpoint directory and the records.

If your proALPHA runs on-premises, the interface must be reachable from outside. That is an arrangement with your IT, not a fault in the connector. In the proALPHA cloud it usually already is.

No. Here even the vendor itself draws the line: the Read API is the reading half of the toolkit, writing runs over completely different paths that anymize does not address. On top comes our own rule that there are no write tools.

The M2M user should still only carry the REST-SERVICES permission. Two boundaries are better than one.

• Filtering by field contents does not work yet. proALPHA filters via parameters named after your own fields, and nobody knows those in advance. The assistant pages instead.
• A page holds at most 50 records. That is on purpose: large extracts through the interface hit your application server; proALPHA itself writes that in its own documentation.
• The field description helps understanding: proALPHA field names are often abbreviated German column names, and the assistant can translate them into readable labels.

proALPHA runs on your own installation, on-premises or in the vendor's cloud. anymize talks only to the address you entered, and the credentials sit encrypted in your workspace.

What comes from the ERP is business data in pure form. Customer names, addresses, contacts, customer and supplier numbers and bank details are replaced by placeholders before a record goes to an AI model.