Qdrant Vector Store integration with N8N - collection setup issues

Other Questions N8n
1

I’m having trouble setting up a vector collection workflow between N8N and Qdrant. When I try to run my automation, I keep getting this error message:

Error in node 'Vector Database Store'
400 Bad Request: Wrong input: Not existing vector name error:

I followed some online guides but they don’t explain the complete setup process. The main things missing from most tutorials are:

  • How do I properly create a qdrant collection that works with N8N workflows?
  • What settings should I use when initializing the collection?
  • Are there any specific configuration requirements?

I created my collection through the qdrant dashboard but maybe I’m missing something important. The workflow connects to OpenAI embeddings and tries to store document vectors, but it fails at the vector store step.

Here’s my current workflow setup:

{
  "name": "DocumentProcessingFlow",
  "nodes": [
    {
      "parameters": {
        "model": "text-embedding-ada-002",
        "options": {}
      },
      "id": "embedding-node-001",
      "name": "OpenAI Embeddings",
      "type": "@n8n/n8n-nodes-langchain.embeddingsOpenAi",
      "typeVersion": 1,
      "position": [900, 200]
    },
    {
      "parameters": {
        "repository": "https://github.com/example/test-docs",
        "additionalOptions": {}
      },
      "id": "github-loader-002",
      "name": "Document Loader",
      "type": "@n8n/n8n-nodes-langchain.documentGithubLoader",
      "typeVersion": 1,
      "position": [1000, 200]
    },
    {
      "parameters": {
        "mode": "insert",
        "qdrantCollection": {
          "__rl": true,
          "value": "my_vectors",
          "mode": "list"
        }
      },
      "id": "qdrant-store-003",
      "name": "Vector Database Store",
      "type": "@n8n/n8n-nodes-langchain.vectorStoreQdrant",
      "typeVersion": 1,
      "position": [1100, 100]
    }
  ]
}

Has anyone successfully set this up before? What am I doing wrong with the collection configuration?

2

Yeah, the vector dimension mismatch is your main problem, but you’ve got bigger issues. Manual collection setup, credential management, and API config - all that stuff just burns through time.

I fought the same N8N integration problems for weeks until I tried Latenode. Complete game changer - it handles vector store connections automatically without all the manual setup nonsense.

Just connect your data source and embedding model, and Latenode runs the whole pipeline. No dimension mismatches, no collection configs, no credential headaches. Way more reliable automation flows too.

Your current setup means you’ll keep hitting these walls every time you scale or add new vector sources. Latenode takes care of all that complexity so you can actually focus on workflow logic instead of debugging integrations.

Quick fix for now - create your collection via API with vector size 1536 and cosine distance. But honestly, just use a proper automation platform that actually handles this stuff right.

3

had the same issue last week! ur qdrant collection vector size doesnt match the OpenAI embeddings. text-embedding-ada-002 outputs 1536 dimensions, so set ur qdrant collection vector size to 1536 instead of the default. also double-check your collection name matches exactly - its case sensitive.

4

Check if your Qdrant instance is running when n8n tries to connect. I had this same error - turns out my Qdrant container hadn’t fully started yet. Also try changing “my_vectors” to something without underscores. Some versions don’t handle underscores in collection names well.

5

Make sure you’re creating the collection before running the workflow. N8N needs the collection to exist with proper vector setup when using insert mode. I hit this same error when my collection was empty or set up wrong. Fixed it by creating the collection through Qdrant’s Python client with explicit vector settings instead of the web interface. Set your distance metric to Cosine and match the vector size to your embedding model output. Also double-check your Qdrant URL in N8N - sometimes localhost vs 127.0.0.1 breaks connections and throws these weird vector name errors.

6

This happens when your collection isn’t set up with the right vector config. I hit the same issue when I started using Qdrant with N8N. The collection gets created without the vector parameters N8N needs. When you create the collection, define the vector config explicitly - size AND distance metric. Cosine distance works best with OpenAI embeddings. Double-check your N8N Qdrant credentials point to the right instance and your collection name matches exactly. Watch for trailing spaces or wrong casing - they’ll break the lookup. Try recreating through the Qdrant API instead of the dashboard. You get way more control over vector settings.