To launch a Phantom from Make with full control over its input, add the PhantomBuster Make an API Call module to your scenario, set the method to POST and the URL to /api/v2/agents/launch, then pass two things in the request body: the Phantom's id, and its argument, which is the Phantom's input in JSON format. This gives you the same control as a direct API call and is ideal for advanced or customized workflows - for a no-code option, use the built-in Launch a Phantom module instead.
Before you start
- You have connected PhantomBuster to Make.
If you haven't, follow How to Connect PhantomBuster with Make first. The connection stores your PhantomBuster API key and handles authentication for you.
→ If you don't have an API key, see Create an API key. API keys can only be created by Workspace admins - if you don't see the Add API key button, check with your admin. - The Phantom you want to launch is already created and set up in your PhantomBuster workspace. With this method you're launching an existing Phantom and sending it the input it needs. You can change that input per run, but the Phantom itself must exist first.
Step 1: Add the "Make an API Call" module
Step 2: Set the endpoint and method
- URL: /api/v2/agents/launch
- Method: POST
Step 3: Add the header
- Key: Content-Type → Value: application/json
You don't need to add your API key as a header, your PhantomBuster connection already authenticates the request for you.
Step 4: Build the request body
The body is JSON with two parts: the Phantom's id, and its argument, which contains the input data.
-
Find your Phantom ID:
Open the Phantom's console in PhantomBuster and copy the number in the browser URL between /phantoms/ and /console.
For more detail, see How to find your Agent ID for API calls. -
Get the exact input for argument:
Go to the Phantom → Setup → the three dots on the input section → Switch to JSON, and copy the JSON shown.
This JSON is the Phantom's expected argument: it shows you exactly which fields that Phantom needs to run. -
For LinkedIn Phantoms and other Phantoms that work on logged-in platforms, the argument must include your session identity: a sessionCookie and a browser userAgent.
Step 5: Run it and check the result
Troubleshooting
- Launch fails or errors: confirm the API key is valid, the Phantom id is correct and in quotes, and the argument JSON is valid.
- LinkedIn Phantom fails: make sure you included the sessionCookie and userAgent. See How to use a different session cookie for a single Phantom launch via API.
- "[404] Agent not found": the Phantom id is wrong or from another Workspace. See How to Fix the "[404] Agent not found" Error.
- JSON errors: a missing bracket, quote, or comma is the usual cause.
Frequently asked questions
How do I launch a PhantomBuster Phantom from Make?
Add the Make an API Call module, POST to /api/v2/agents/launch, and pass the Phantom's id and argument as JSON in the body.
When should I use "Make an API Call" instead of the "Launch a Phantom" module?
Use Make an API Call when you need full control over the launch input or want to customize the request. For a simpler setup, the built-in Launch a Phantom module works.
Do I need to add my PhantomBuster API key as a header in Make?
No. Your PhantomBuster connection in Make authenticates the request automatically. The only header you need is Content-Type: application/json.
Where do I find my Phantom ID?
In the Phantom's console URL: the number between /phantoms/ and /console.
Why does my LinkedIn Phantom fail to launch from Make?
It's missing the session identity. Include a sessionCookie and userAgent in the argument.
Can I launch a Workflow from Make?
No. Workflows can't be launched via API. Chain individual Phantoms instead.