Exporting all Architect flows as JSON via CX as Code CLI

General Discussion
1

Could someone explain the correct syntax for exporting all Architect flows as JSON using the CX as Code CLI? I am trying to automate this for our nightly backups.

  1. Run nice-cxone export --flow
  2. Receive Error: No flows found. Use --all to export all.
  3. Run nice-cxone export --flow --all
  4. Output is empty or partial.

The documentation is vague on the --all flag behavior with specific resource types. Is there a hidden parameter for bulk export?

2

nice-cxone export --flow --all --destination ./exports is correct, but you likely hit the 500-flow pagination limit or lack the architect:flow:read scope. Use nice-cxone export --flow --all --destination ./exports --verbose to check for 403s, or switch to the Python SDK’s get_architect_flow with a cursor loop for reliable bulk extraction.

3

Pretty sure the CLI often fails on bulk exports due to scope limitations. The docs state “architect:flow:read” is required, but you also need “architect:flow:export”. Here is the minimal payload structure the CLI expects for validation.

{
 "type": "flow",
 "exportAll": true,
 "scope": ["architect:flow:read", "architect:flow:export"]
}
4

Yep, this is a known issue… the CLI silently drops flows with invalid JSON references during bulk export. Do not rely on --all for critical backups. Use the REST API directly to ensure integrity.

curl -X GET "https://api.mypurecloud.com/api/v2/architect/flows" \
 -H "Authorization: Bearer <token>" \
 -H "Accept: application/json" | jq '.entities[] | {id, name}'
5

As far as I remember, the cli export is just a thin wrapper around the standard rest endpoint, and it often fails to handle the cursor pagination correctly for large environments. you are hitting that 500-flow limit mentioned above, but the real issue is usually the token expiration during long-running bulk jobs. instead of relying on the cli flags, use the python sdk to handle the retry logic and token refresh automatically. here is a quick snippet to get all flows with proper cursor handling:

from platformclientv2 import Configuration, ArchitectApi, context

config = Configuration()
config.host = 'api.mypurecloud.com'
context.set_default(config)
architect_api = ArchitectApi()

flows = []
cursor = None
while True:
 resp = architect_api.get_architect_flows(page_size=1000, cursor=cursor)
 flows.extend(resp.entities)
 if not resp.next_uri:
 break
 cursor = resp.next_uri.split('cursor=')[1]

this ensures you get every single flow without silent drops. the json structure remains consistent, so you can dump it directly to a file.