This workflow follows the Agent → Chat Trigger recipe pattern — see all workflows that pair these two integrations.
The workflow JSON
Copy or download the full n8n JSON below. Paste it into a new n8n workflow, add your credentials, activate. Full import guide →
{
"name": "[Vorlage] Agent \u2014 RAG (Qdrant Wissensbasis)",
"nodes": [
{
"parameters": {
"content": "## [Vorlage] RAG-Agent (Qdrant)\n\n**Vor dem ersten Lauf:**\n1. Ollama-Credential anlegen: Base URL `http://llm-service:11434`, in beiden Ollama-Nodes ausw\u00e4hlen.\n2. Qdrant-Credential anlegen: *QdrantApi* \u2192 URL `http://qdrant:6333` (API-Key leer lassen).\n3. Im Node \u00bbQdrant Wissensbasis (Tool)\u00ab den Collection-Namen `MEINE_COLLECTION` durch eine echte Collection ersetzen.\n4. Embedding-Modell ziehen: `docker exec llm-service ollama pull bge-m3`.\n\n**WICHTIG \u2014 Embedding-Kompatibilit\u00e4t:**\nDie Suche funktioniert nur gegen Collections, deren Vektoren mit **demselben** Embedding-Modell erzeugt wurden wie hier konfiguriert (`bge-m3` via Ollama, 1024 Dimensionen). Die Plattform-Collection des Dokument-Indexers nutzt BGE-M3 \u00fcber den separaten embedding-service \u2014 gleiche Modellfamilie/Dimension, aber vor produktiver Nutzung gegen die eigene Collection verifizieren (Score-Stichprobe). F\u00fcr eine eigene Wissensbasis: Collection mit diesem Workflow-Setup bef\u00fcllen (Insert-Mode des Qdrant-Nodes) \u2014 dann passt es garantiert.\n\nDetails: docs/integrations/N8N_AGENTS.md",
"height": 560,
"width": 440,
"color": 5
},
"id": "sticky-setup-rag",
"name": "Sticky Note \u2014 Setup",
"type": "n8n-nodes-base.stickyNote",
"typeVersion": 1,
"position": [
-180,
40
]
},
{
"parameters": {
"options": {}
},
"id": "chat-trigger-rag",
"name": "Beim Chat-Empfang",
"type": "@n8n/n8n-nodes-langchain.chatTrigger",
"typeVersion": 1.1,
"position": [
320,
240
]
},
{
"parameters": {
"options": {
"systemMessage": "Du bist ein Assistent mit Zugriff auf eine lokale Wissensbasis (Qdrant-Vektordatenbank auf einer Arasul-Appliance, alles on-device).\n\nDein Werkzeug:\n- wissensbasis_suche: semantische Suche in der Wissensbasis. \u00dcbergib eine pr\u00e4zise, eigenst\u00e4ndige Suchanfrage (keine Pronomen, keine Bez\u00fcge auf den Chatverlauf).\n\nArbeitsweise:\n1. Suche IMMER zuerst in der Wissensbasis, bevor du antwortest; formuliere die Suchanfrage ggf. um und suche erneut, wenn die Treffer nicht passen.\n2. Antworte NUR auf Basis der gefundenen Passagen. Wenn die Wissensbasis nichts Passendes enth\u00e4lt, sage das ehrlich \u2014 erfinde keine Inhalte.\n3. Nenne am Ende kurz, auf welche Fundstellen (Metadaten/Quelle der Treffer) sich deine Antwort st\u00fctzt.\n4. Antworte auf Deutsch."
}
},
"id": "agent-rag-1",
"name": "RAG-Agent",
"type": "@n8n/n8n-nodes-langchain.agent",
"typeVersion": 2.2,
"position": [
620,
240
]
},
{
"parameters": {
"model": "qwen3:8b",
"options": {
"temperature": 0.2,
"numCtx": 32768
}
},
"id": "ollama-lm-rag",
"name": "Ollama Chat Model (qwen3:8b)",
"type": "@n8n/n8n-nodes-langchain.lmChatOllama",
"typeVersion": 1,
"position": [
500,
480
],
"credentials": {
"ollamaApi": {
"name": "<your credential>"
}
}
},
{
"parameters": {
"mode": "retrieve-as-tool",
"toolName": "wissensbasis_suche",
"toolDescription": "Semantische Suche in der lokalen Wissensbasis (Qdrant). Eingabe: eine pr\u00e4zise, eigenst\u00e4ndige Suchanfrage in nat\u00fcrlicher Sprache. Liefert die relevantesten Textpassagen mit Metadaten zur\u00fcck.",
"qdrantCollection": {
"__rl": true,
"value": "MEINE_COLLECTION",
"mode": "id"
},
"topK": 4,
"options": {}
},
"id": "qdrant-tool-1",
"name": "Qdrant Wissensbasis (Tool)",
"type": "@n8n/n8n-nodes-langchain.vectorStoreQdrant",
"typeVersion": 1.3,
"position": [
760,
480
],
"credentials": {
"qdrantApi": {
"name": "<your credential>"
}
}
},
{
"parameters": {
"model": "bge-m3"
},
"id": "ollama-embed-1",
"name": "Ollama Embeddings (bge-m3)",
"type": "@n8n/n8n-nodes-langchain.embeddingsOllama",
"typeVersion": 1,
"position": [
760,
700
],
"credentials": {
"ollamaApi": {
"name": "<your credential>"
}
}
}
],
"connections": {
"Beim Chat-Empfang": {
"main": [
[
{
"node": "RAG-Agent",
"type": "main",
"index": 0
}
]
]
},
"Ollama Chat Model (qwen3:8b)": {
"ai_languageModel": [
[
{
"node": "RAG-Agent",
"type": "ai_languageModel",
"index": 0
}
]
]
},
"Qdrant Wissensbasis (Tool)": {
"ai_tool": [
[
{
"node": "RAG-Agent",
"type": "ai_tool",
"index": 0
}
]
]
},
"Ollama Embeddings (bge-m3)": {
"ai_embedding": [
[
{
"node": "Qdrant Wissensbasis (Tool)",
"type": "ai_embedding",
"index": 0
}
]
]
}
},
"active": false,
"settings": {
"executionOrder": "v1",
"saveManualExecutions": true,
"callerPolicy": "workflowsFromSameOwner"
},
"id": "arasul-vorlage-agent-rag",
"meta": {
"templateCredsSetupCompleted": false
},
"tags": []
}
Credentials you'll need
Each integration node will prompt for credentials when you import. We strip credential IDs before publishing — you'll add your own.
ollamaApiqdrantApi
For the full experience including quality scoring and batch install features for each workflow upgrade to Pro
About this workflow
[Vorlage] Agent — RAG (Qdrant Wissensbasis). Uses chatTrigger, agent, lmChatOllama, vectorStoreQdrant. Chat trigger; 6 nodes.
Source: https://github.com/koljaschoepe/arasul-jet/blob/602cf1858b1c6b560d06417cb593792eb85b5141/services/n8n/templates/agents/agent-rag.json — original creator credit. Request a take-down →
Related workflows
Workflows that share integrations, category, or trigger type with this one. All free to copy and import.
Provides one workflow to maintain the knowledge base and another one to query the knowledge base. Uploaded documents are saved into the Qdrant vector store. When a query is made, the most relevant doc
RAG Agent Integration Hub mit Knowledge Management. Uses memoryPostgresChat, lmChatOllama, lmOllama, toolVectorStore. Chat trigger; 27 nodes.
Apple RAG Chatbot V2. Uses googleDrive, documentDefaultDataLoader, agent, toolVectorStore. Chat trigger; 23 nodes.
Local RAG AI Agent with Knowledge Management. Uses memoryPostgresChat, lmChatOllama, lmOllama, toolVectorStore. Chat trigger; 22 nodes.
⚠️ Note: This system only works for self-hosted n8n instances. It will not function on n8n.cloud or other remote setups. LocalRAG.AI is a private, on-prem AI assistant that uses your own documents to