Version 0.1.0-beta.6

TL;DR: Create segmentations directly, and patch a single scenario node without resending the whole graph. Scenario and banner search gain real filtering and pagination, so large projects stop timing out. Search report tools now accept a domain key on its own for sites with no view configured.

Create segmentations directly

• Create a new segmentation with one or more named segments, each defined by its own customer filter. An optional global filter pre-filters customers before segment assignment, so anyone who doesn't match it is excluded from every segment.
• Fetch a similar segmentation first with search_segmentations and adapt its segments and filter — the most reliable way to get valid conditions for a project's schema.

Tags can't be set on create. Apply them from the dashboard afterward.

Patch a single scenario node

• Change fields on one existing scenario node without resending the whole flow, including any email or SMS design payloads on other nodes.
• The node is fetched, your patch is merged onto it, and the full scenario is saved — every other node, connection, and untouched field on the target node is preserved.

Nested fields such as a node's customer filter, event, or design are replaced wholesale, not merged. Sending a partial one is rejected with a preview of what it would delete, unless you confirm the deletion is intentional. Adding or removing nodes, or editing connections, still goes through the existing scenario-update tool.

One tool for exploring scenario nodes

The two tools for mapping a scenario's nodes and fetching one in full are now a single tool: omit the node ID for a lightweight map of the whole graph, or pass it for one node's complete configuration. The scenario is cached for 10 seconds, so back-to-back calls against the same scenario share a single request.

Filter and page through scenarios and banners

• Scenario search now filters by status, tags, creation date, trigger type, and more, and pages through results with offset and limit.
• Banner search filters by status, serving state, tags, and more, with the same pagination.
• Both default to live records only, matching the dashboard, and archived records are opt-in.

Neither tool paginated before, so listing everything — including archived records — was the only option, and it timed out on large projects. Filtering and paging through results avoids that.

Bug fixes

• Search report tools that accept a domain key now work when no view is configured for that site — passing the domain key alone no longer silently falls back to account-level metrics.
• Search accounts in the EU are now routed to the EU cluster instead of the US one, fixing errors on merchandising and analytics calls for those accounts.
• The scenario node cache lasts 10 seconds, not 60 as previously documented.