Process models (BPMN) · Exported file link

BPMN 2.0 files and TestingOps: scenario candidates from any modeling tool

If your process models come from a tool without its own connection, a BPMN 2.0 export is enough. Link the .bpmn file, and TestingOps reads the process with the same parser it uses for Camunda or SAP Signavio models.

Exported file link
Available

Request a demo BPMN 2.0 specification (OMG)

BPMN 2.0 file File TestingOpsAI test agents

Connection: Exported file link
Writes back: Nothing, the file is only read
Last checked: 7 Nov 2026

In short

BPMN 2.0 file and TestingOps in short

The BPMN 2.0 file connection reads a .bpmn file exported from any modeling tool, from a link to a file in GitHub or GitLab or at any other https address; the file is linked, not uploaded.

The file must be BPMN 2.0 XML with <definitions> in the BPMN namespace, otherwise it is refused with a clear message, and each process becomes a knowledge-base document with steps, decisions, exceptions, form fields and numbered paths from start to end.

Paths are bounded to 30 per process and 60 steps per path, with loops followed once and parallel branches listed separately, and they are candidates for review rather than finished scenarios.

What it does

What it does for your testing

1
Any tool's export becomes test input
Input
A .bpmn file exported from your modeling tool
TestingOps
Checks that it is BPMN 2.0 and parses each process with the parser the direct connections use
Output
A process document in the project's knowledge base, whichever tool drew the model
2
Paths become a review list
Input
Start events, gateways, sequence flows and end events
TestingOps
Walks from each start event and numbers each path, with flow names or conditions in brackets
Output
A bounded list of scenario candidates to keep, merge or drop
3
The committed file drives updates
Input
A file kept in a Git repository
TestingOps
Hashes the file on each sync and re-reads it only when its content changed
Output
Candidates that follow the committed model

Step by step

How it works

Export BPMN 2.0Export the process from your modeling tool as a BPMN 2.0 file (.bpmn) and keep it where TestingOps can fetch it, for example in a Git repository.
Add the BPMN 2.0 file connectionUnder Integrations → Platforms → Process models (BPMN 2.0) → BPMN 2.0 file, save the connection. Add an access token only for private links; it is sent as a bearer token.
Link the fileIn the project's Knowledge tab, add a source under Process models (BPMN 2.0) and paste the file's link. GitHub and GitLab page links become raw file links, and diagrams.net links that point to a file are followed.
Sync and reviewSync now reads the file; automatic syncs skip it while it is unchanged. Agents and testers read the process documents and write scenarios from the candidates they keep.

Connect

Setup

Export the process from your modeling tool as BPMN 2.0 and commit the .bpmn file to a repository, or publish it at an https address TestingOps can reach.
In TestingOps, open Integrations → Platforms → Process models (BPMN 2.0) → BPMN 2.0 file and save the connection, with an access token only for private links.
In the project's Knowledge tab, select Add source under Process models (BPMN 2.0) and paste the file's link.
Select Sync now. Commit a new version to the same path and the next sync reads it.

Documentation: BPMN 2.0 specification (OMG)

Where you connect it in TestingOps: Integrations → Platforms → Process models (BPMN 2.0) → BPMN 2.0 file

Use cases

Use models from a tool without a connector

Trigger: A team keeps its processes in a modeling tool that exports BPMN 2.0
Uses: The exported file at a stable link
Agent activity: Reads the process document and writes scenarios from the candidates
Result: Scenario drafts without moving the models to another tool
Condition: The export must be BPMN 2.0 XML; other formats are refused

Follow a model kept in Git

Trigger: A model change is merged into the branch the link points to
Uses: The file's content on the next sync
Agent activity: Works from the updated paths when writing or revising scenarios
Result: Candidates that match the merged model
Condition: Only the file at the link is read; link the branch you want, such as main

Cover sub-processes in one read

Trigger: A model with embedded sub-processes and transactions
Uses: Inner steps on the same path with the sub-process name as a prefix, and start events inside sub-processes
Agent activity: Writes scenarios that include the inner steps
Result: Scenarios that read end to end
Condition: A call activity is one step; the process it calls is not followed

Capabilities

Capabilities and limits

What is read

A BPMN 2.0 XML file (<definitions> in the BPMN 2.0 namespace, with the bpmn:, bpmn2: or no prefix) from a link: GitHub and GitLab page links become raw file links, diagrams.net links that point to a file are followed, other http(s) addresses are fetched as they are
Per process (named after its pool when there is one): start, end, intermediate and boundary events with their type, such as timer, error or message; tasks, including user, service, script, business rule, manual, send and receive tasks; call activities, without following the called process; sub-processes, ad-hoc sub-processes and transactions with their inner steps; exclusive, inclusive, parallel, event-based and complex gateways; sequence flows with their name or condition; lanes; documentation; inline form fields such as camunda:formField
Not parsed: message flows between pools, data objects and data stores, text annotations, the DMN tables behind business rule tasks, linked forms, and timer or expression details beyond the event type

What is produced

One knowledge-base document per process: steps with kind, lane and documentation, decisions with their outcomes, exceptions from boundary events, form fields, and numbered paths from start to end marked as test scenario candidates

Where it can write

Nothing: the file is only downloaded and read

How it connects

A link to the file, with an optional access token for private links (sent as a bearer token). There is no live connection to the modeling tool

Account and permissions

No account in any modeling tool
Read access to the link; an access token only when the link is private

Limits

Paths are bounded: at most 30 per process and 60 steps per path; a loop is followed once and the path ends with "back to" the step it returns to; each outgoing branch of a gateway, parallel and inclusive ones included, becomes its own path, so branches are not combined
Paths are candidates for review: no TestingOps scenarios or cases are created from them automatically
Files that are not BPMN 2.0 are refused; Google Drive links are refused because they need a Google sign-in
The file is read from its link; uploading a .bpmn file is not supported
Up to 12,000 characters per process section and 200 processes per model; files and exports up to 15 MB

Example

Example: how a team uses it

Illustrative scenario, not a customer story.

An architecture team keeps its claim intake process in a modeling tool without its own TestingOps connection, exports it as BPMN 2.0 and commits the file to GitHub next to its test plans. The document lists two exclusive gateways and a timer boundary event on the Request missing documents task.

The paths show five candidates, one of them ending in "back to Check claim" where a loop returns. The team drops one duplicate, and an agent writes scenarios for the other four.

Frequently asked questions

Which modeling tools does this work with?

Any tool whose export is BPMN 2.0 XML with <definitions> in the BPMN 2.0 namespace, whatever the prefix. Other formats are refused with a message. Tools with their own connection, such as Camunda or SAP Signavio, can also be read directly.

Why a link and not an upload?

A link lets TestingOps fetch the file again on every sync and notice when it changed, so the candidates follow your model. Uploading BPMN files is not supported.

Can the file be private?

Yes. Add an access token to the connection and it is sent as a bearer token when the file is fetched, for example from a private GitHub or GitLab repository. Google Drive links are refused because they need a Google sign-in.

Do the paths become test scenarios on their own?

No. Up to 30 paths per process and 60 steps per path are listed as candidates; a loop is followed once and each parallel branch is its own path. Nothing becomes a TestingOps scenario or case until an agent or a tester writes it.

What happens to call activities and sub-processes?

A call activity appears as one step and the process it calls is not followed. Embedded sub-processes, ad-hoc sub-processes and transactions are expanded, with their inner steps prefixed by the sub-process name.

Next step

Bring one exported .bpmn file to a demo and see the scenario candidates TestingOps finds in it.

Request a demo More process models integrations