Automate with n8n or Make

Use n8n or Make HTTP requests to turn sheets and forms into Krevaya drafts, send draft digests, and track publishing results

Written by the Krevaya team · Updated

n8n and Make work with Krevaya through their generic HTTP request modules and a Krevaya API key. Krevaya doesn't currently offer a published n8n node, Make app, or importable workflow templates, so every recipe below uses plain HTTP requests.

Before you start

  • API access is included with Plus and up.
  • Base URL: https://api.krevaya.com. Send the key as the header Authorization: Bearer YOUR_API_KEY, stored in n8n's or Make's credential store. Never put it in a URL, a spreadsheet, or a workflow note.
  • Keep every ID as text. Krevaya IDs can be larger than a spreadsheet or JavaScript number can hold exactly.

Create the right key

Open Settings → API keys & MCP.

  • For drafts only: expand Connecting with Make? and choose Prepare Make key (it works for n8n too), or pick the Drafts for review preset. Then select Create API key and copy it. This key can read and create drafts. It can't approve, schedule, publish, generate AI content, or upload media.
  • For reporting only: choose Read-only connection or the Read only preset.
  • To schedule posts or turn links into drafts: add Create & publish posts or Generate content under Adjust individual permissions.

Recipe: sheet, form, or CRM to a draft for review

  1. Call GET /v1/automation/brands once and copy your brand's id.
  2. For each ready record, send POST /v1/automation/drafts with Content-Type: application/json and an Idempotency-Key header set to a stable ID for that record, such as content-sheet:row-42:linkedin:v1 (8 to 200 characters: letters, numbers, and . _ : / -).

{

  "brand_id": "YOUR_BRAND_ID",

  "platform": "linkedin",

  "body": "A useful update for our community.",

  "asset_ids": []

}

  1. Save the returned id on your record.

A new draft returns 201. Repeating the same request with the same key returns 200 and the header Idempotency-Replayed: true, without a second draft. The same key with different content returns 409, and a key whose draft was deleted returns 410. Use a new key only when you mean to create another draft. These drafts always wait for a person, even when the brand's Publish mode is Auto.

Recipe: morning draft digest

Trigger on a schedule, then call GET /v1/automation/posts?brand_id=YOUR_BRAND_ID&status=draft&limit=25. Send each post's id, platform, and body to your email or chat step. If next_cursor isn't null, pass it as before to get the next page.

Recipe: publishing results to a calendar or alert

Poll GET /v1/automation/events?brand_id=YOUR_BRAND_ID&after=0&limit=100. Each event has an event_id, a type (post.published or post.failed), and the post's post_id, platform, permalink_url, and error.

  • Process the whole page, then store next_cursor and send it as after on the next run.
  • Upsert by event_id so a repeat never adds a second row or alert.
  • Events are kept for 30 days. A 410 means your cursor expired: reconcile with the posts list, then start again from after=0.

For instant delivery, use webhooks instead. n8n's Webhook node and Make's webhooks don't check Krevaya's signature on their own, so verify X-Krevaya-Signature against the raw body before acting on an event.

Recipe: schedule an approved draft

With Create & publish posts, POST /v1/posts/{id}/schedule with {"schedule_at": "2026-12-01T15:00:00Z", "approve": true} approves and schedules a draft. Treat approve: true as the reviewer's decision, not a default. Pinterest and Tumblr drafts are refused: a person approves those in Krevaya.

With Generate content, POST /v1/generate with {"url": "https://example.com/article", "brand_id": "YOUR_BRAND_ID"} returns post_ids once the drafts exist. It can take a minute or two, so raise the HTTP timeout, and it uses AI credits. Turn off automatic retries for this step: a repeat creates more drafts and spends credits again.

Limits and errors

  • Your account shares 120 requests per minute across every key, REST, MCP, and GraphQL. On 429, wait for Retry-After.
  • 401 means reconnect with a valid key. 403 means check the key's permissions and your plan. 422 names the field to fix, including text over a channel's character limit.

Did this answer your question?