Set up n8n credentials

filtrat works in n8n with the standard HTTP Request node. There's no community node to install. You store your API key once as a credential, and every filtrat node uses it.

1. Create an API key

1. In filtrat, open API → API keys.
2. Press Create key and name it n8n.
3. Copy the key. It starts with ak_ and is shown only once.

2. Create the credential in n8n

1. In n8n, open Credentials and create a new credential of type Header Auth.
2. Fill it in:

3. Save it.

The header value is the word Bearer, one space, then your full key.

3. Use it in an HTTP Request node

Every filtrat call in n8n uses the same node settings:

For example, to verify an email coming from an earlier node, set Specify Body to Using JSON, switch the JSON field to Expression and paste:

{{ JSON.stringify({ email: $json.email, resolve_catchall: false }) }}

Set the node's Timeout option to 90 seconds or more if you turn on resolve_catchall or call the finder: these can take up to a minute.

4. Test the credential

Add an HTTP Request node with method GET, URL https://api.filtrat.io/v1/credits/balance and the Filtrat API credential, then run it. It's free. You should see your balance. A 401 means the header value is wrong: check it reads Bearer, a space, then the full key.

Calling filtrat from other automation tools

The same three things work in Make, Zapier, Pipedream or your own scripts:

• the URL https://api.filtrat.io/v1/...;
• the header Authorization: Bearer ak_your_key;
• a JSON body for POST requests.

Set the request timeout to 90 seconds and retry on 429, 502, 503 and 504. A 429 carries a Retry-After header with the seconds to wait. The single verify and find endpoints allow 200 requests a minute per account; Rate limits lists every endpoint. The API quick start shows the two most common calls, and the full API reference lists the rest.

Good practice

• Use a separate key for n8n, so you can revoke it without breaking other tools.
• For lists, start a bulk job instead of calling the single endpoint in a loop. It's faster and stays well within your rate limit.
• Handle errors on the node: with On Error → Continue, one bad row doesn't stop the whole run.