Skip to content
Echopoint Echopoint Docs

Echopoint AI

Echopoint AI is a chat copilot built into the flow canvas. Open the panel next to the flow you are working on and ask it to inspect the flow or propose the next change.

What it can do

Section titled “What it can do”

The copilot works on the flow you have open:

  • Inspect — ask what the flow does, which assertions cover a node, or where a variable like {{payment_id}} comes from.
  • Modify — ask for a concrete change: “add an assertion that $.status equals succeeded on the create-payment node”, “add a Delay node before verify-refund”, “rename these nodes after what they produce”.

It speaks the same vocabulary as the canvas: Request, Delay, and Module nodes, success and error edges, outputs, assertions.

Every AI change is restorable

Section titled “Every AI change is restorable”

When Echopoint AI applies a change, it does so with a restorable snapshot. If the result is not what you wanted, restore the snapshot and the flow returns to its state before that AI run — a built-in undo for AI edits.

That makes the copilot safe to use on flows that matter:

  1. Ask for a change.
  2. Review what was applied on the canvas.
  3. Keep it, or restore the snapshot to undo the AI run.

Working with it effectively

Section titled “Working with it effectively”
  • Be specific. “Add a statusCode equals 201 assertion to create-payment” lands better than “make this more robust”.
  • Review before running. AI edits show up on the canvas like any other edit — read the nodes and assertions it touched before you press Run flow.
  • Validate after structural changes. echopoint flows validate <flow-id> confirms edges and bindings still resolve.
  • Publish what you approve. Once an AI-assisted edit is reviewed and green, publish a version so CI keeps running the snapshot you signed off on.

Next steps

Section titled “Next steps”