Make.com Integration

This guide walks you through adding FlautoPsy webhook monitoring to your Make.com workflows.

Step 1: Get Your Webhook URL

  1. Log in to FlautoPsy at https://flautopsy.stralocroft.com/dashboard
  2. Go to Settings → Setup (or /setup)
  3. Copy your unique webhook URL (looks like
    text
    https://flautopsy.stralocroft.com/api/webhook?workspace_id=...
    )

Step 2: Add an HTTP Module to Your Workflow

  1. Open your Make.com workflow
  2. Click the + button to add a new step
  3. Search for and select the HTTP module (under "Apps")
  4. Click Make a request

Step 3: Configure the HTTP Module

Set the following in your HTTP module:

| Field | Value | |-------|-------| | URL | Paste your FlautoPsy webhook URL from Step 1 | | Method | POST | | Content-Type | application/json |

Headers

Add the header:

  • Key:
    text
    Content-Type
  • Value:
    text
    application/json

Step 4: Map Your Workflow Data

Under Request body, paste the following JSON structure:

json
{
  "workflow_id": "your-workflow-name",
  "prompt": "<<prompt_from_your_llm>>",
  "output": "<<output_from_your_llm>>",
  "latency_ms": <<latency_in_milliseconds>>,
  "cost_tokens": <<total_tokens_used>>,
  "model": "gpt-4"
}

For each field, map it to the corresponding output from your LLM module:

  • workflow_id: A fixed identifier for this workflow (e.g.,
    text
    "email-classifier"
    )
  • prompt: The prompt you sent to the LLM (use bundle mapping:
    text
    {{1.prompt}}
    or similar)
  • output: The LLM's response (use bundle mapping:
    text
    {{1.response}}
    )
  • latency_ms: How long the LLM call took in milliseconds
  • cost_tokens: Total tokens used (prompt + completion)
  • model: The model name (e.g.,
    text
    "gpt-4"
    ,
    text
    "claude-3"
    )

Example for OpenAI Module

If your workflow uses the OpenAI module, your HTTP body might look like:

json
{
  "workflow_id": "email-classifier",
  "prompt": "{{1.messages[0].content}}",
  "output": "{{1.choices[0].message.content}}",
  "latency_ms": {{1.duration}},
  "cost_tokens": {{add(1.usage.prompt_tokens; 1.usage.completion_tokens)}},
  "model": "gpt-4"
}

Replace

text
1
with the correct module number in your workflow.

Step 5: Test the Integration

  1. Click Save and then Run once on the HTTP module
  2. Check FlautoPsy dashboard → Traces to see if the trace arrived
  3. Run your workflow 10+ times to build a baseline

Common Gotchas

Content-Type Header Missing

If your traces aren't being recorded, ensure the

text
Content-Type: application/json
header is set. Make.com sometimes omits this.

Incorrect Field Mapping

  • latency_ms must be a number, not a string
  • cost_tokens must be a number
  • prompt and output can be truncated to 10,000 characters if very long

Workflow ID Not Set

Use a consistent

text
workflow_id
for all traces from the same workflow. If you change it, FlautoPsy will treat it as a new workflow.

Troubleshooting

"Webhook request failed"

  • Verify your webhook URL is correct (copy again from /setup)
  • Check that your Make.com workflow has internet access
  • Test by sending a manual HTTP request with Postman or curl

"I sent traces but they're not showing up"

  • Check the FlautoPsy dashboard → Traces tab
  • Verify the JSON is valid (all numbers are numbers, not strings)
  • Ensure your workspace has at least 10 traces before alerts trigger

How do I delete my workflow from FlautoPsy?

Go to FlautoPsy dashboard → Workflows → click your workflow → Delete. This removes all associated traces.

Next Steps