What's new in Echopoint

Every flow now has a run history. Open any run to see per-node results, every recorded assertion, and the exact step that caused a failure. Node-level history shows how a single step has behaved across runs.

When an upstream step fails, downstream nodes are skipped with a recorded reason and the list of inputs they were missing. Steps marked run_when: always still execute, so cleanup keeps running after a failure.

Tag your flows and run them as a suite: echopoint flows run --tag smoke locally, or pass tags to the GitHub Action with match-mode any or all. Flows are searchable by tag and full text in the app.

Every assertion now records expected, actual, and passed on each run — including passing ones. When a run goes green, you can still see exactly what was checked.

Use {{$uuid}}, {{$email}}, {{$fullName}}, {{$creditCard}} and 65 more in any request. Values are faker-backed and deterministically seeded per execution, so each run's data is reproducible.

The CLI is public: a checksum-verified installer (no sudo), browser-based login, and full command groups for flows, collections, and environments. nanostack-dev/echopoint-cli@v1 runs flows from CI on an ephemeral runner and reports per-flow results.

Launch a flow and watch it run: node-by-node progress streams to the canvas over SSE — flow.started, node.started, node.completed, node.failed — with no refresh.

Upload an OpenAPI spec and get a collection of ready-to-run requests, organized in folders. Pair it with OpenAPI Sync to keep the collection reconciled as the spec changes.

Ask Echopoint AI to inspect a flow or propose the next change, right in the canvas. Every AI change is applied with a restorable snapshot, so anything it does can be undone.

Publishing a flow creates an immutable version. Browse and restore previous snapshots, and pin a launch to a specific version-id so CI always runs exactly what you reviewed.