Make

Create a key in Cloud Platform → API → API Keys (starts with px_live_). Find the workflow id in the Visual Studio URL.

• URL: https://www.prompteryx.ai/api/v1/workflows/WORKFLOW_ID/execute, Method: POST
• Headers: Authorization = Bearer px_live_YOUR_KEY
• Body type: Raw, Content type: JSON (application/json), Request content:

{
  "variables": { "postcode": "{{1.postcode}}" },
  "options": { "runMode": "headless" },
  "webhookUrl": "https://hook.eu1.make.com/YOUR_CUSTOM_WEBHOOK"
}

• Turn on Parse response. The run id is data.execution.id.
• webhookUrl is optional: leave it out and poll instead (below), or point it at a Make custom webhook to get the result pushed (section 3).

Poll for the result

Add Tools → Sleep (15 s), then HTTP → Make a request: GET https://www.prompteryx.ai/api/v1/executions/{{2.data.execution.id}} with the same header and Parse response on. Use a Repeater with a filter, or Make's break/retry error handler, until data.execution.status is COMPLETED or FAILED. Step results are in data.execution.outputs, one entry per step name.

HTTP → Make a request, POST https://www.prompteryx.ai/api/v1/ai-browser/execute, same header, JSON body:

{
  "instruction": "Go to https://news.ycombinator.com and return the titles of the top 5 stories",
  "maxSteps": 20,
  "mode": "start"
}

Then repeat a second request to the same URL until data.terminal is true:

{ "jobId": "{{2.data.jobId}}", "mode": "run" }

• Set the module's Timeout to 300 seconds; each call works for up to about 4 minutes.
• The answer is data.finalAnswer. data.jobStatus is done, failed, or awaiting_input when the agent stopped to ask before a risky step.
• Runs spend AI credits and cloud minutes; cap them with maxSteps and maxCredits. See Run the AI browser agent.

1. Start a scenario with Webhooks → Custom webhook, click Add, name it and copy the address.
2. Put that address in webhookUrl when you start a run (section 1).
3. Run once so Make learns the data structure: it receives { event, data: { executionId, workflowId, status, outputs, error, duration } }.

Each callback is signed (X-Prompteryx-Signature-256, HMAC-SHA256 with your whsec_ secret). To check it in Make, enable Get request headers on the webhook and compare with the sha256 function as described in Webhooks, or keep the webhook address private and treat it as the secret.

Add a Webhook Trigger node to the Prompteryx workflow (authentication: for example a secret in X-Webhook-Secret) and save. Then in Make, HTTP → Make a request:

POST https://www.prompteryx.ai/api/v1/webhooks/WORKFLOW_ID
Header  X-Webhook-Secret: your_workflow_secret
Body    { "email": "{{1.email}}", "action": "process" }

1. Add MCP Client → Call a tool (or, in an AI agent, Add tool → MCP Client).
2. Create a connection: + New MCP server, URL https://www.prompteryx.ai/api/mcp.
3. In API Key / Access token paste your px_live_ key.
4. Pick a tool, for example workflow_list, workflow_execute or browser_ai_action, and map its arguments.

Make's MCP Client is in open beta on Make's side. Tool reference: MCP tools.