Fix: n8n HighLevel (GoHighLevel) node not working

    The native n8n HighLevel node drops custom fields, fails after a green credential test, and misses half the API. Here's the HTTP Request pattern that replaces it reliably.

    Last updated 2026-11-10~6 min read
    n8nGoHighLevelHTTP RequestPrivate Integration

    Symptoms — what you're seeing

    • The n8n HighLevel node runs green but custom fields on the contact never change
    • The credential test passes, then real operations fail with 401, 403 or a generic node error
    • Fields, pipelines or operations you need are missing from the node's dropdowns, or the UI shows the wrong options
    • A workflow that used the node fine stopped working after a GHL or n8n update
    • An AI Agent using the HighLevel tool fails while the same call works elsewhere

    Why this happens

    Short version: it isn't you. The native HighLevel node has been unreliable for a long time. There are open issues on the n8n repository for the node's UI (#13489), inconsistent credentials when used with HTTP nodes (#12230) and general node errors (#12689). The community thread on updating custom fields through it sits in n8n's Known Issues category and runs to a second page. n8n's own guidance for unsupported operations is the HTTP Request node.

    There are three underlying problems. The node covers only part of the GHL v2 API, so anything outside that (custom objects, many calendar and conversation operations) isn't there. The way it sends custom fields doesn't always match what the API expects, so the request succeeds and the field silently stays empty. And GHL has moved from v1 API keys to OAuth and Private Integrations, which left older credentials and older node versions in a half-working state.

    Waiting for an upstream fix isn't a plan for a production client workflow. The fix that works today is to replace the node with HTTP Request calls you control.

    The fix, step by step

    1. Create a Private Integration token in the sub-account. In the client's sub-account: Settings → Private Integrations → Create new integration. Grant only what the workflow needs, for example contacts.readonly, contacts.write and locations/customFields.readonly.
    2. Make one reusable Header Auth credential. Header name Authorization, value Bearer pit-.... Every HTTP Request node below uses it, plus a Version: 2021-07-28 header and the base URL https://services.leadconnectorhq.com.
    3. Fetch the custom field IDs once. The API wants field IDs, not the labels you see in the UI:
      GET /locations/{locationId}/customFields
      → { "customFields": [ { "id": "aBc123...", "name": "Lead Source Detail", "fieldKey": "contact.lead_source_detail" }, ... ] }
      Store the ID-to-name map in a Set node or a static data table so the workflow reads by name.
    4. Upsert contacts with custom fields in one call. Upsert creates or updates by email/phone, so retries don't create duplicates:
      POST /contacts/upsert
      {
        "locationId": "{{ $json.locationId }}",
        "email": "{{ $json.email }}",
        "firstName": "{{ $json.firstName }}",
        "customFields": [
          { "id": "aBc123...", "field_value": "{{ $json.leadSourceDetail }}" }
        ]
      }
      Check the response: the returned contact's customFields should contain your value. If it doesn't, the ID is wrong or the value doesn't fit the field type (dropdown values must match an option exactly).
    5. Rebuild the other operations the same way. Update a contact with PUT /contacts/{contactId}, add tags with POST /contacts/{contactId}/tags, create opportunities with POST /opportunities/, book with POST /calendars/events/appointments. Put each one in a small sub-workflow ("GHL – Upsert Contact", "GHL – Add Tag") and call it with Execute Workflow, so every client workflow uses the same tested building block.
    6. Add retries and a read-back. On each HTTP Request node, turn on Retry On Fail (3 tries, with a wait). For critical writes, follow with a GET and compare. That's what catches the silent no-op the native node gave you.

    Common variations

    • "Custom field update returns 200 but nothing changes" — you're sending the field's name or key instead of its ID, or a value that doesn't match a dropdown option. Use the ID from /customFields.
    • "The credential test passes but operations fail" — the OAuth connection was authorised against a different sub-account, or it's an old API-key credential. Recreate it while logged into the right sub-account, or switch to the HTTP Request pattern.
    • "I get 'The token does not have access to this location'" — token and location don't match. See GHL token does not have access to this location.
    • "The HighLevel tool fails inside the AI Agent" — wrap the HTTP Request sub-workflow as a tool with a clear description and typed inputs instead. The agent gets a predictable interface and you get the logs.
    • "Calendar requests 404 with 'resource could not be found'" — the calendar ID belongs to another sub-account, or the endpoint is wrong. List calendars for the location first.

    When this isn't the issue

    Want a node that behaves? The n8n custom node TypeScript boilerplate walks through building your own, and the nodes directory lists the ones I maintain.

    Browse 4,900+ ready-to-import workflow templates →

    Need this workflow customized or debugged? Book a 30-min scope call →

    Stuck configuring this workflow?

    I build and fix n8n workflows for agencies and SaaS teams. 30-minute scope call — bring the problem, leave with an estimate.

    Related pages

    Related workflows