You shouldn't need to write code to build a production AI agent. Today, you don't have to.
We're launching the AACFlow Visual DAG Workflow Editor — a fully visual, drag-and-drop canvas for building AI agents. Built on ReactFlow. Compiles directly into our DAG execution engine. Designed for operators, business analysts, and domain experts — not just developers.
What It Looks Like
The editor is a canvas. You drag blocks from a sidebar — "Gmail: Read Inbox," "OpenAI: Chat Completion," "AmoCRM: Create Deal," "PostgreSQL: Run Query," "Conditional Branch" — and drop them onto the canvas. You connect them with edges. Each connection defines the flow of data: this block's output feeds that block's input.
When you select a block, a side panel opens with its configuration. For the Gmail block: which mailbox, what search query, how many emails to fetch. For the OpenAI block: which model, what system prompt, what temperature. For the Conditional block: what variable to check, what comparison operator, what branches to activate.
That's it. No code. No YAML. No JSON. Just blocks, edges, and configuration.
What Happens Under the Hood
When you click "Save," the visual graph doesn't stay visual. It compiles into an optimized DAG execution plan:
- Parse — the editor serializes the graph: all nodes, their types, their configurations, and all edges connecting them.
