Apify Website Crawler → Qdrant Sync with OpenAI Embeddings (n8n)

Keep your Qdrant knowledge base fresh—automatically sync Apify crawls with OpenAI embeddings in n8n

This n8n workflow triggers when an Apify Website Content Crawler run finishes, then syncs the crawl dataset into your Qdrant collection using OpenAI embeddings. It also produces a freshness report showing vectors before/after and which chunks were added, refreshed, or deleted.

What this workflow does

• Triggers on completion: Starts when an Apify Website Content Crawler actor run is completed.
• Captures a baseline: Calls the Qdrant Collections API to read the current point count in the target collection (vectors-before).
• Performs the sync: Launches an Apify Qdrant Integration actor run to upsert the crawl dataset into Qdrant, generating OpenAI embeddings and applying delta updates plus deletion of expired chunks.
• Waits for settling: Pauses for 30 seconds to allow the Qdrant sync to settle.
• Verifies the result: Calls the Qdrant Collections API again to read the post-sync point count (vectors-after).
• Reports freshness: Fetches the Apify actor run logs and outputs a report with vectors-before/after and counts of chunks added, refreshed, and deleted.

Use cases

• SaaS operators maintaining an up-to-date semantic search or RAG index from a website.
• Automation engineers who want automated vector delta updates (including deletion of expired chunks) after crawling.
• Teams using Qdrant + OpenAI embeddings that need visibility into indexing changes after every crawl.

Technical details

• n8n nodes / tools: Code, Wait, Sticky Note, HTTP Request, apify/n8n-nodes-apify (Apify Trigger + Apify runs)
• External APIs: Qdrant Collections API (point count reads) and Apify actor run logs
• Environment variables: QDRANT_API_KEY and OPENAI_API_KEY (ensure env access is enabled in nodes if required)
• Setup requirement: Replace https://YOUR-CLUSTER.qdrant.io:6333 with your Qdrant endpoint and ensure the target collection name (e.g., docs) is correct or allowed to auto-create.