Skip to main content
The Prompt-to-Flow AI Architect transforms natural language intent into fully connected, executable Quixa workflow graphs grounded against your active API Collections.

​
1. How It Works

  1. On the empty canvas or from the top toolbar, click the AI Architect button (✨).
  2. Describe your target business orchestration in plain English:
    “Authenticate against the user service, retrieve the user’s latest orders, loop through each order to check shipment status, and if an order is delayed, trigger an email notification.”
  3. The AI Architect engine performs a 3-step synthesis:
    • Semantic Intent & Spec Grounding: Matches your prompt against registered OpenAPI endpoints and collections.
    • DAG Topology Construction: Synthesizes the logical sequence of HTTP nodes, For loops, Switch routers, and Stop nodes.
    • Variable Lineage Synthesis: Automatically binds predecessor properties ({{AuthStep.response.body.token}}) into downstream headers and parameters.
  4. Click Apply to Canvas to render the connected graph immediately.

​
2. Example Prompts

  • Authentication & Ingestion:
    “Login with credentials from environment variables, fetch pending invoices, and POST each invoice to our analytics warehouse.”
  • Error Handling & Multi-Branch Fallbacks:
    “Call the payment gateway. If status is 200, log order success. If status is 402 or 400, route to customer notification and terminate the branch.”
  • Synthetic Data Generation:
    “Create 10 synthetic user accounts with randomized names, emails, and phone numbers, and verify they exist in the search index.”