Helix workflow engine for governed enterprise operations Learn more →

Visual Editor

Draw the flow. Run it in production.

The Visual Editor is the canvas for designing multi-agent A2A flows. Draw the graph: agents as nodes, handoffs as edges, Cortex knowledge queries as inputs. The resulting flow is a durable, auditable workflow — not a prototype that breaks in production.

Studio & Grid →
Features

Visual design with production-grade execution.

Unlike flowchart tools that generate pseudocode, the Visual Editor generates real executable workflows. What you draw is what runs.

Agent graph canvas

Place agent nodes on a canvas, draw handoff edges, and configure the Cortex knowledge queries each agent uses as context. The visual graph is the executable flow — no separate configuration layer.

Live blast-radius overlay

Activate the blast-radius overlay to see which downstream services each agent step could affect. High-risk paths are highlighted before you execute — not after.

Consent-model visualisation

Each edge in the visual editor carries a consent scope. The editor shows which data categories cross which agent boundaries — making the data consent model visible at design time.

Export as YAML or API call

Every flow in the editor exports as a YAML workflow definition or a single POST to /v1/workflows. Version-control your agent flows alongside your code.

Reusable agent templates

Save frequently-used sub-graphs as templates. Share them across teams in the Studio library. Every template carries the same RBAC and audit constraints as a hand-written workflow.

Step-level Cortex bindings

For each agent step, bind the Cortex entities it reads and writes. The editor validates the binding against the live knowledge graph schema — catching entity name typos before a run is launched.

Design + execution gap

Not a diagram tool. The actual workflow.

Canvas to runtime in one step

No intermediate code generation step. No "export and configure separately." The canvas definition is the runtime definition — parsed directly by the Cendriix orchestrator.

Blast-radius at design time

Risk is visible on the canvas, not discovered in a post-incident review. High-risk edges are colour-coded by blast-radius score before you hit run.

Versioned alongside your code

YAML export means agent flows live in your repo. PRs, code review, and CI checks apply to workflow changes the same way they apply to application code.

Enterprise governance built in

Consent scopes, RBAC policies, and cost caps are first-class properties of every edge and node in the Visual Editor — not an afterthought.

Design your first multi-agent flow.

Visual canvas. Blast-radius overlay. Production-grade execution.