Système RAG vectoriel PDF : Mistral OCR, base de données Weaviate et serveur MCP
Intermédiaire
Ceci est unDocument Extraction, Multimodal AIworkflow d'automatisation du domainecontenant 13 nœuds.Utilise principalement des nœuds comme Set, MistralAi, FormTrigger, McpTrigger, RerankerCohere. Construire un système de recherche PDF avec Mistral OCR et la base de données Weaviate
Prérequis
- •Aucun prérequis spécial, prêt à l'emploi après importation
Nœuds utilisés (13)
Aperçu du workflow
Visualisation des connexions entre les nœuds, avec support du zoom et du déplacement
Exporter le workflow
Copiez la configuration JSON suivante dans n8n pour importer et utiliser ce workflow
{
"id": "rV1w47cZn1rsk7MP",
"meta": {
"instanceId": "402614e91d3374d08756ba129f83d1678d5ada16d0573bfa1398e795ef5f7589",
"templateCredsSetupCompleted": true
},
"name": "PDF to Vector RAG System: Mistral OCR, Weaviate Database and MCP Server",
"tags": [],
"nodes": [
{
"id": "d9e90589-d6b6-4601-bac8-5009b765fa78",
"name": "Cohere Embeddings",
"type": "@n8n/n8n-nodes-langchain.embeddingsCohere",
"position": [
160,
336
],
"parameters": {
"modelName": "embed-multilingual-v3.0"
},
"typeVersion": 1
},
{
"id": "5e7c6668-64a4-4cc2-b519-ab75f07ecab5",
"name": "Chargeur de documents",
"type": "@n8n/n8n-nodes-langchain.documentDefaultDataLoader",
"position": [
-144,
336
],
"parameters": {
"options": {},
"textSplittingMode": "custom"
},
"typeVersion": 1.1
},
{
"id": "b808993d-a6b9-497f-88b5-271c16abc185",
"name": "Cohere Reranker",
"type": "@n8n/n8n-nodes-langchain.rerankerCohere",
"position": [
304,
336
],
"parameters": {},
"typeVersion": 1
},
{
"id": "117948ee-4be7-4869-b35b-d0c58a66fcd5",
"name": "Serveur de connaissances MCP",
"type": "@n8n/n8n-nodes-langchain.mcpTrigger",
"position": [
192,
-80
],
"webhookId": "4f3bebe7-f088-4131-b3ad-c4e300f74517",
"parameters": {
"path": "c74c97f5-0197-45e3-b4dd-f3efbd4bab22",
"authentication": "headerAuth"
},
"typeVersion": 2
},
{
"id": "700afe32-2bcc-4f31-a680-cdce710861e2",
"name": "Rechercher la base de connaissances",
"type": "@n8n/n8n-nodes-langchain.vectorStoreWeaviate",
"position": [
256,
128
],
"parameters": {
"mode": "retrieve-as-tool",
"options": {},
"useReranker": true,
"toolDescription": "Use this tool to search and retrieve information from the knowledge base containing various documents and resources",
"weaviateCollection": {
"__rl": true,
"mode": "list",
"value": "KnowledgeDocuments",
"cachedResultName": "KnowledgeDocuments"
},
"includeDocumentMetadata": false
},
"typeVersion": 1.3
},
{
"id": "4075cf1b-e9f6-44cc-b827-29fa5eb1ee97",
"name": "Instructions de téléchargement",
"type": "n8n-nodes-base.stickyNote",
"position": [
-880,
-16
],
"parameters": {
"color": 5,
"width": 688,
"height": 304,
"content": "## Manual Document (PDF) Upload Section\nThis section allows users to upload PDF files to the knowledge base. The files will be processed by Mistrals OCR and stored in the vector database for later retrieval and search."
},
"typeVersion": 1
},
{
"id": "bf763fa9-03f4-4010-a6a9-63b6bb104fa1",
"name": "Séparateur de texte",
"type": "@n8n/n8n-nodes-langchain.textSplitterRecursiveCharacterTextSplitter",
"position": [
-144,
480
],
"parameters": {
"options": {
"splitCode": "markdown"
},
"chunkSize": 600,
"chunkOverlap": 200
},
"typeVersion": 1
},
{
"id": "d1fe1c74-d9bc-4040-8446-47e90365c5f7",
"name": "Télécharger PDF",
"type": "n8n-nodes-base.formTrigger",
"position": [
-832,
112
],
"webhookId": "ff827d70-a596-4f81-a99e-63bf42189fd0",
"parameters": {
"options": {
"ignoreBots": true,
"buttonLabel": "Upload Document",
"appendAttribution": true
},
"formTitle": "Upload Documents to Knowledge Base",
"formFields": {
"values": [
{
"fieldType": "file",
"fieldLabel": "PDF File",
"multipleFiles": false,
"requiredField": true,
"acceptFileTypes": ".pdf"
}
]
},
"responseMode": "lastNode",
"formDescription": "Upload PDF files to the knowledge base for AI-powered search and retrieval"
},
"typeVersion": 2.2
},
{
"id": "1f04ecce-0277-4a88-9514-fec6b45ba1cf",
"name": "Extraire le texte du PDF",
"type": "n8n-nodes-base.mistralAi",
"position": [
-608,
112
],
"parameters": {
"options": {},
"binaryProperty": "file"
},
"retryOnFail": true,
"typeVersion": 1
},
{
"id": "53c82051-ce9b-432e-b90e-d5eb83483e49",
"name": "Préparer les données documentaires",
"type": "n8n-nodes-base.set",
"position": [
-384,
112
],
"parameters": {
"options": {},
"assignments": {
"assignments": [
{
"id": "518ae17b-b486-4438-8151-f49afb3b68eb",
"name": "filename",
"type": "string",
"value": "={{ $('Upload PDF').item.json.file.filename }}"
},
{
"id": "a574ee4d-6341-4fd5-ac8e-9452eff70aa1",
"name": "content",
"type": "string",
"value": "={{ $json.extractedText }}"
},
{
"id": "metadata-source",
"name": "source",
"type": "string",
"value": "uploaded_pdf"
},
{
"id": "metadata-timestamp",
"name": "upload_timestamp",
"type": "string",
"value": "={{ new Date().toISOString() }}"
}
]
}
},
"typeVersion": 3.4
},
{
"id": "d465c50a-87e9-4824-83ca-d5662630590c",
"name": "Stocker dans la base de données vectorielle",
"type": "@n8n/n8n-nodes-langchain.vectorStoreWeaviate",
"position": [
-112,
112
],
"parameters": {
"mode": "insert",
"options": {},
"weaviateCollection": {
"__rl": true,
"mode": "list",
"value": ""
}
},
"typeVersion": 1.3
},
{
"id": "5a5e4028-3764-4bb2-8900-df77c0c47bde",
"name": "Note adhésive",
"type": "n8n-nodes-base.stickyNote",
"position": [
112,
304
],
"parameters": {
"color": 4,
"width": 320,
"height": 288,
"content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n## Embedding and Rerank\nYou can exchange the models, but you **must** use the same model for embedding and retrieval and **no switching** later on"
},
"typeVersion": 1
},
{
"id": "f0ecfbb5-e9ce-49fd-a921-672b12b7ef13",
"name": "Note adhésive1",
"type": "n8n-nodes-base.stickyNote",
"position": [
112,
-240
],
"parameters": {
"width": 336,
"height": 288,
"content": "## MCP Server Trigger\nYou can call this MCP Server as a tool in your AI Workflow"
},
"typeVersion": 1
}
],
"active": false,
"pinData": {},
"settings": {
"executionOrder": "v1"
},
"versionId": "8b3d2e49-d84f-4fca-bdda-298f3da3c15b",
"connections": {
"d1fe1c74-d9bc-4040-8446-47e90365c5f7": {
"main": [
[
{
"node": "1f04ecce-0277-4a88-9514-fec6b45ba1cf",
"type": "main",
"index": 0
}
]
]
},
"bf763fa9-03f4-4010-a6a9-63b6bb104fa1": {
"ai_textSplitter": [
[
{
"node": "5e7c6668-64a4-4cc2-b519-ab75f07ecab5",
"type": "ai_textSplitter",
"index": 0
}
]
]
},
"b808993d-a6b9-497f-88b5-271c16abc185": {
"ai_reranker": [
[
{
"node": "700afe32-2bcc-4f31-a680-cdce710861e2",
"type": "ai_reranker",
"index": 0
}
]
]
},
"5e7c6668-64a4-4cc2-b519-ab75f07ecab5": {
"ai_document": [
[
{
"node": "d465c50a-87e9-4824-83ca-d5662630590c",
"type": "ai_document",
"index": 0
}
]
]
},
"d9e90589-d6b6-4601-bac8-5009b765fa78": {
"ai_embedding": [
[
{
"node": "700afe32-2bcc-4f31-a680-cdce710861e2",
"type": "ai_embedding",
"index": 0
},
{
"node": "d465c50a-87e9-4824-83ca-d5662630590c",
"type": "ai_embedding",
"index": 0
}
]
]
},
"1f04ecce-0277-4a88-9514-fec6b45ba1cf": {
"main": [
[
{
"node": "53c82051-ce9b-432e-b90e-d5eb83483e49",
"type": "main",
"index": 0
}
]
]
},
"53c82051-ce9b-432e-b90e-d5eb83483e49": {
"main": [
[
{
"node": "d465c50a-87e9-4824-83ca-d5662630590c",
"type": "main",
"index": 0
}
]
]
},
"700afe32-2bcc-4f31-a680-cdce710861e2": {
"ai_tool": [
[
{
"node": "117948ee-4be7-4869-b35b-d0c58a66fcd5",
"type": "ai_tool",
"index": 0
}
]
]
}
}
}Foire aux questions
Comment utiliser ce workflow ?
Copiez le code de configuration JSON ci-dessus, créez un nouveau workflow dans votre instance n8n et sélectionnez "Importer depuis le JSON", collez la configuration et modifiez les paramètres d'authentification selon vos besoins.
Dans quelles scénarios ce workflow est-il adapté ?
Intermédiaire - Extraction de documents, IA Multimodale
Est-ce payant ?
Ce workflow est entièrement gratuit et peut être utilisé directement. Veuillez noter que les services tiers utilisés dans le workflow (comme l'API OpenAI) peuvent nécessiter un paiement de votre part.
Workflows recommandés
RAG de documentation et chatbot basé sur Weaviate et OpenAI
Chatbot de questions-réponses pour documents RAG : interroger le contenu de PDF avec Weaviate et OpenAI
Set
Form Trigger
Extract From File
+
Set
Form Trigger
Extract From File
17 NœudsMary Newhauser
Extraction de documents
Centre de commandement IA commercial : Agent modulaire Google Workspace, recherche vectorielle et rapports multicanal
Centre de commande commercial AI : Agents modulaires Google Workspace, recherche vectorielle et rapports multi-canaux
Set
Gmail
Slack
+
Set
Gmail
Slack
80 NœudsPaul
Extraction de documents
Analyse de actions par IA basée sur Danelfin, TwelveData et Alpha Vantage
Système d'analyse boursière par IA intégrant Danelfin, TwelveData et Alpha Vantage
Set
Code
Gmail
+
Set
Code
Gmail
74 NœudsPaul
Trading crypto
Automatisation du développement de partenaires commerciaux avec Google Maps, GPT-4 et WhatsApp
Automatisation du développement de partenaires commerciaux avec Google Maps, GPT-4 et WhatsApp
If
Set
Code
+
If
Set
Code
64 NœudsKhairul Muhtadin
Chatbot IA
Reclassement #1
Automatisation du marketing téléphonique avec le web scraping, GPT-4o et WhatsApp
Set
Code
Webhook
+
Set
Code
Webhook
48 NœudsKhaisa Studio
Nurturing de leads
Automatisation du trading d'actions pilotée par l'IA
Automatiser le trading d'actions avec l'analyse technique par IA et la négociation Alpaca
Set
Code
Gmail
+
Set
Code
Gmail
96 NœudsPaul
Trading crypto
Informations sur le workflow
Niveau de difficulté
Intermédiaire
Nombre de nœuds13
Catégorie2
Types de nœuds10
Description de la difficulté
Auteur
Dietmar
@docdMedical specialist in internal medicine, gastroenterology, and infectious diseases. Building innovative healthcare automation workflows with n8n, integrating AI, speech-to-text, and medical data standards for efficient clinical documentation and analysis.
Liens externes
Voir sur n8n.io →
Partager ce workflow