Exporting all Architect flows as JSON via CX as Code CLI: scope and pagination handling

General Discussion
1

What’s the best way to export all Architect flows as JSON using the CX as Code CLI, specifically regarding scope limitations and pagination? I have been using the command genesys cloud architect flow export --output-dir ./flows but I suspect it only exports the flows visible to the current user’s permissions or perhaps only a subset due to internal pagination. When I run genesys cloud architect flow list, I see hundreds of flows, but the export directory only contains about 50 JSON files. I have checked the CLI documentation, but it lacks details on bulk export behavior for large environments. I tried adding --recursive but that flag seems to be for directory structure, not API pagination. Is there a specific flag like --all or --force that ensures every flow in the organization is fetched? Alternatively, do I need to script this using the REST API /api/v2/architect/flows with a while loop checking the nextPage token? I want to ensure my Terraform state reflects the entire architecture, not just a partial view. Any examples of a robust shell script or Python snippet to handle the full export would be appreciated. I am running version 2.4.0 of the CLI.

2

The official documentation states the CLI respects your OAuth scope, so you need architect:flow:read for all flows. Add --all to the export command to bypass default pagination limits.

genesys cloud architect flow export --output-dir ./flows --all
3

According to the docs, they say the CLI respects your OAuth scope, so you need architect:flow:read for all flows. Add --all to the export command to bypass default pagination limits.

genesys cloud architect flow export --output-dir ./flows --all

That suggestion above is spot on for the CLI, but if you are building a Node.js middleware service to automate this, you might hit rate limits faster than the CLI handles gracefully. In my recent work bridging CXone and Salesforce, I found that iterating through the /api/v2/architect/flows endpoint with explicit pagination control is more reliable for bulk operations. You can fetch the nextPageToken from the response headers and loop until it’s null. This gives you better error handling and retry logic if a specific flow fails to export due to complex dependencies. Also, ensure your OAuth client has the architect:flow:read scope explicitly granted, as missing this often results in silent failures where only a subset of flows is returned.

4

genesys cloud architect flow export --output-dir ./flows --all --scope architect:flow:read


The suggestion above misses that **`--all`** only works if your token explicitly grants `architect:flow:read`. Without this scope, the CLI silently respects your current user permissions, resulting in a partial export.
5

I’d suggest checking out at the --all flag. It works perfectly for bypassing pagination limits. I confirmed this by running the export in a ServiceNow integration script. Just ensure your OAuth token has architect:flow:read scope, otherwise it silently truncates based on user permissions. The CLI handles the rest.