n8n, Zapier & CI

There is no Premation node to install and no SDK to pull in. Every one of these is a plain HTTP request, which is the point — the API was shaped around what a generic HTTP node can already do.

Three design choices exist specifically for this:

• X-API-Key is accepted alongside Authorization, because it is the default header of n8n’s generic credential type.
• Download answers 302 rather than streaming bytes, because HTTP nodes follow redirects natively.
• Rate limiting answers 429 with retryAfterSeconds, because that is what automation tools back off on.

Sending a file: upload it, don’t host it

image and media inputs take a URL, and in n8n you usually have the opposite — a PNG sitting in $binary.data from a form trigger, a Drive node or a mail attachment.

POST /v1/assets takes the bytes and hands back a URL:

Method: POST
URL: {{ $env.PREMATION_BASE }}/v1/assets
Authentication: Predefined ▸ Header Auth ▸ Premation API
Send Body: ✓   Body Content Type: n8n Binary File
Input Data Field Name: data

{
  "url": "https://res.cloudinary.com/…/a1b2c3.png",
  "bytes": 284119,
  "mime": "image/png",
  "expiresAt": "2026-08-28T09:00:00.000Z"
}

Put {{ $json.url }} straight into the next node’s inputs. Scope: renders:write, which is in the default grant.

Or point at a URL you already have

If the asset is already public, skip the upload and pass the URL directly. A signed URL works provided it is valid when the render is created — the asset is fetched at job creation, not at render time. A URL needing a header or a cookie will not work: the fetcher sends neither.

n8n

{
  "templateId": "tpl_7f8a9b",
  "inputs": {
    "headline": "{{ $json.title }}",
    "character": "{{ $json.imageUrl }}"
  },
  "callbackUrl": "{{ $execution.resumeUrl }}"
}

Polling instead

If a tunnel is out of the question, drop the callbackUrl and loop:

1. HTTP Request — GET {{ $env.PREMATION_BASE }}/v1/renders/{{ $json.jobId }}
2. IF — is status one of completed, failed, cancelled?
3. Wait 5 seconds on the false branch, then back to step 1.

Remember each poll spends one API request from your monthly quota.

Downloading the file

HTTP Request, GET {{ $env.PREMATION_BASE }}/v1/renders/{{ $json.jobId }}/download, Response Format File. Redirect following is on by default, so the node ends up holding the MP4 as binary data.

Zapier

Zapier has no long-running step, so the webhook path is the only sane one.

GitHub Actions

A nightly render, start to finish:

name: Render weekly reel
 
on:
  schedule: [{ cron: '0 6 * * 1' }]
  workflow_dispatch:
 
jobs:
  render:
    runs-on: ubuntu-latest
    steps:
      - name: Queue the render
        id: queue
        env:
          KEY: ${{ secrets.PREMATION_API_KEY }}
          BASE: ${{ vars.PREMATION_BASE }}
        run: |
          job=$(curl -sS -X POST "$BASE/v1/renders" \
            -H "Authorization: Bearer $KEY" \
            -H "Content-Type: application/json" \
            -H "Idempotency-Key: gha-${{ github.run_id }}" \
            -d '{
              "templateId": "tpl_7f8a9b",
              "inputs": { "headline": "Week ${{ github.run_number }}" }
            }')
          echo "id=$(echo "$job" | jq -r .jobId)" >> "$GITHUB_OUTPUT"
 
      - name: Wait for it
        id: wait
        env:
          KEY: ${{ secrets.PREMATION_API_KEY }}
          BASE: ${{ vars.PREMATION_BASE }}
        run: |
          for _ in $(seq 1 120); do
            job=$(curl -sS "$BASE/v1/renders/${{ steps.queue.outputs.id }}" \
              -H "Authorization: Bearer $KEY")
            status=$(echo "$job" | jq -r .status)
            echo "status=$status"
            case "$status" in
              completed) echo "url=$(echo "$job" | jq -r .videoUrl)" >> "$GITHUB_OUTPUT"; exit 0 ;;
              failed|cancelled) echo "$job" | jq -r .error >&2; exit 1 ;;
            esac
            sleep 10
          done
          echo "Timed out waiting for the render." >&2
          exit 1
 
      - name: Download
        run: curl -sSL -o reel.mp4 "${{ steps.wait.outputs.url }}"
 
      - uses: actions/upload-artifact@v4
        with:
          name: reel
          path: reel.mp4

Make.com and everything else

Any tool that can send an HTTP request with a header works the same way:

1. POST <base>/v1/renders with the key in Authorization.
2. Either supply a callbackUrl your tool can receive on, or poll GET <base>/v1/renders/{jobId} until status is terminal.
3. Use videoUrl, or GET …/download for the bytes.

That is the entire integration surface. See the endpoint reference for the rest.