For development and enterprise IT
The interface – what it does and how to call it.
TAOM makes processes, attributes and functions available through an interface. This page describes what it is for and what an access token looks like. It does not replace a full reference – customers receive that with their access.
Basics
Four things up front
Versioned
The interface sits under a versioned namespace. A new version does not silently replace the old one; existing integrations keep working.
Named access tokens
One token per connecting application, with scope, tenant and expiry date. Revocable individually, usage is recorded.
Scope instead of full access
Read only, read and write, or full access. Deleting requires the highest level. The scope is checked on every call.
Tied to a tenant
If a token carries a tenant, it reaches only that tenant’s workspaces – whatever identifier is sent along.
Authentication
One header field, nothing more
GET https://your-environment/wp-json/taom/v1/… Authorization: Bearer <your access key>
The key is shown once when the token is created and afterwards kept only as a checksum. It belongs on the server, never in a browser application.
Areas
What you can address
Workspaces and models
Create, read, change, fetch versions. Plus phases, metadata, groups of several diagrams and the provenance seal – every model file generated carries it, and it can be verified without access to TAOM.
Enrichment
Start a run, query its state, fetch the result – for a single element or the whole model.
Events from Jira and Confluence
Status changes in Jira and changed text sections in Confluence report back to the model through our own Atlassian app. Which side owns which value is defined field by field. Outbound, return and field ownership →
Attachments
Upload, list, fetch and remove documents – bound to the element they belong to.
Agents
Directory, style profile, library, plus request and result of a run.
Users and licence
Sign-in, account, licence state – the basis for another application acting on behalf of an authorised user.
Presence
Who is working on a model right now. This lets collaboration be reflected outside the interface too.
Two patterns
What an integration usually looks like
Pull
The other application asks, TAOM answers. Suitable for reports, overviews and anything that wants to read the current state.
Job and callback
A run is triggered, TAOM answers immediately with an acknowledgement and reports the result later to a registered address. Necessary because an agent run takes longer than a request may wait.
Recently possible
Query across all workspaces
Querying across the whole inventory works: by attribute, with free text or through thirteen ready-made questions – for instance “inspection required but no test case”. The answer names model, phase and role per hit.
Not there yet
What we should honestly add
Suggestions while modelling
The library from the Learning Model is in place. Offering it as a suggestion while typing is implemented in one place only.
Rate limits
Per-token limits are planned but not yet set.