Supabase Insertion & Upsertion & Retrieval
Efficiently manage and query your data with the Supabase Insertion & Upsertion & Retrieval workflow, a powerful solution for integrating document management with intelligent data processing. This 21-node workflow, triggered manually, connects Google Drive, Supabase, and OpenAI to automate the ingestion, updating, and retrieval of information. It allows you to upload documents from Google Drive, which are then processed by a Recursive Character Text Splitter and embedded using OpenAI Embeddings for insertion or upsertion into your Supabase vector store via the Insert Documents and Update Documents nodes. When a chat message is received, the workflow leverages OpenAI's Chat Model and a Question and Answer Chain to retrieve relevant information from Supabase using the Retrieve by Query node, providing intelligent responses based on your stored documents. This workflow is ideal for businesses and individuals who need to maintain an up-to-date knowledge base, power AI-driven chatbots with proprietary information, or automate the synchronization of document content with a searchable database, significantly reducing manual data entry and improving information accessibility.
Workflow JSON
{"meta": {"instanceId": "1a23006df50de49624f69e85993be557d137b6efe723a867a7d68a84e0b32704"}, "nodes": [{"id": "54065cc9-047c-4741-95f6-cec3e352abd7", "name": "Google Drive", "type": "n8n-nodes-base.googleDrive", "position": [2700, -1840], "parameters": {"fileId": {"__rl": true, "mode": "url", "value": "https://drive.google.com/file/d/xxxxxxxxxxxxxxx/view"}, "options": {}, "operation": "download"}, "typeVersion": 3}, {"id": "62af57f5-a001-4174-bece-260a1fc595e8", "name": "Default Data Loader", "type": "@n8n/n8n-nodes-langchain.documentDefaultDataLoader", "position": [3120, -1620], "parameters": {"loader": "epubLoader", "options": {}, "dataType": "binary"}, "typeVersion": 1}, {"id": "ce3d9c7c-6ce9-421a-b4d0-4235217cf8e6", "name": "Sticky Note", "type": "n8n-nodes-base.stickyNote", "position": [2620, -2000], "parameters": {"width": 749.1276349295781, "height": 820.5109034066329, "content": "# INSERTING\n\n- it's important to use the same embedding model when for any interaction with your vector database (inserting, upserting and retrieval)"}, "typeVersion": 1}, {"id": "81cb3d3e-70af-46c8-bc18-3d076a222d0b", "name": "Sticky Note1", "type": "n8n-nodes-base.stickyNote", "position": [1720, -1160], "parameters": {"color": 3, "width": 873.9739981925188, "height": 534.0012007720542, "content": "# UPSERTING\n"}, "typeVersion": 1}, {"id": "60ebdb71-c7e0-429b-9394-b680cc000951", "name": "Sticky Note2", "type": "n8n-nodes-base.stickyNote", "position": [1720, -2000], "parameters": {"color": 4, "width": 876.5116990000852, "height": 821.787041589866, "content": "# PREPARATION (in Supabase)\n\n- your database needs the extension 'pgvector' enabled -> select Database > Extension > Search for 'vector'\n- make sure you have a table that has the following columns (if not, use the query below in the Supabase SQL Editor)\n\n```\nALTER TABLE \"YOUR TABLE NAME\"\nADD COLUMN embedding VECTOR(1536), // check which number of dimensions you need (depends on the embed model)\nADD COLUMN metadata JSONB,\nADD COLUMN content TEXT;\n```\n\n- make sure you have the right policies set -> select Authentication > Policies\n- make sure you have the custom function `match_documents` set up in Supabase -> This is needed for the Vector Store Node (as query name) \n(if not, use the query below in the Supabase SQL Editor to create that function)\n- make sure you check the size of the AI model as it should be the same vector size for the table \n(e.g. OpenAI's Text-Embedding-3-Small uses 1536)\n\n```\nCREATE OR REPLACE FUNCTION public.match_documents(\n filter JSONB,\n match_count INT,\n query_embedding VECTOR(1536) // should match same dimensions as from insertion\n)\nRETURNS TABLE (\n id BIGINT,\n content TEXT,\n metadata JSONB,\n embedding VECTOR(1536), // should match same dimensions as from insertion\n similarity FLOAT\n)\nLANGUAGE plpgsql AS $$\nBEGIN\n RETURN QUERY\n SELECT\n v.id,\n v.content,\n v.metadata,\n v.embedding,\n 1 - (v.embedding <=> match_documents.query_embedding) AS similarity\n FROM \"YOUR TABLE NAME\" v\n WHERE v.metadata @> filter\n ORDER BY v.embedding <=> match_documents.query_embedding\n LIMIT match_count;\nEND;\n$$\n;\n```\n"}, "typeVersion": 1}, {"id": "ae95b0c3-b8b3-44eb-8070-b1bc6cac5cd2", "name": "Sticky Note3", "type": "n8n-nodes-base.stickyNote", "position": [3400, -2000], "parameters": {"color": 5, "width": 810.9488123113013, "height": 821.9537074055816, "content": "# RETRIEVAL"}, "typeVersion": 1}, {"id": "58168721-cbd7-498c-9d16-41b4d5c6a68f", "name": "Question and Answer Chain", "type": "@n8n/n8n-nodes-langchain.chainRetrievalQa", "position": [3680, -1860], "parameters": {}, "typeVersion": 1.3}, {"id": "ddf1228f-f051-445b-8a42-54c2510a0b2e", "name": "OpenAI Chat Model", "type": "@n8n/n8n-nodes-langchain.lmChatOpenAi", "position": [3600, -1680], "parameters": {"options": {}}, "typeVersion": 1}, {"id": "734a2c48-b445-4e62-99b7-dc1dcd921c52", "name": "Vector Store Retriever", "type": "@n8n/n8n-nodes-langchain.retrieverVectorStore", "position": [3760, -1680], "parameters": {"topK": 10}, "typeVersion": 1}, {"id": "43f761b7-f4da-4b29-8099-9b2c15f79fe9", "name": "Recursive Character Text Splitter1", "type": "@n8n/n8n-nodes-langchain.textSplitterRecursiveCharacterTextSplitter", "position": [3120, -1460], "parameters": {"options": {}}, "typeVersion": 1}, {"id": "de0d2666-88e4-4a4d-ba46-cf789b9cba85", "name": "Customize Response", "type": "n8n-nodes-base.set", "notes": "output || text", "position": [4020, -1860], "parameters": {"options": {}, "assignments": {"assignments": [{"id": "440fc115-ccae-4e30-85a5-501d0617b2cf", "name": "output", "type": "string", "value": "={{ $json.response.text }}"}]}}, "notesInFlow": true, "typeVersion": 3.4}, {"id": "a396671f-a217-4f05-b969-cb64f10e4b01", "name": "When chat message received", "type": "@n8n/n8n-nodes-langchain.chatTrigger", "position": [3480, -1860], "webhookId": "d7431c58-89aa-4d70-b5bd-044be981b3a9", "parameters": {"public": true, "options": {"responseMode": "lastNode"}, "initialMessages": "=Hi there! \ud83d\ude4f\n\nYou can ask me anything about Venerable Geshe Kelsang Gyatso's Book - 'How To Transform Your Life'\n\nWhat would you like to know? "}, "typeVersion": 1.1}, {"id": "6312f6bc-c69c-4d4f-8838-8a9d0d22ed55", "name": "Retrieve by Query", "type": "@n8n/n8n-nodes-langchain.vectorStoreSupabase", "position": [3700, -1520], "parameters": {"options": {"queryName": "match_documents"}, "tableName": {"__rl": true, "mode": "list", "value": "Kadampa", "cachedResultName": "Kadampa"}}, "typeVersion": 1}, {"id": "ba6b87b9-e96d-47a3-83f8-169d7172325a", "name": "Embeddings OpenAI Retrieval", "type": "@n8n/n8n-nodes-langchain.embeddingsOpenAi", "position": [3700, -1360], "parameters": {"options": {}}, "typeVersion": 1}, {"id": "bcd1b31f-c60b-4c40-b039-d47dadc86b23", "name": "Embeddings OpenAI Insertion", "type": "@n8n/n8n-nodes-langchain.embeddingsOpenAi", "position": [2920, -1620], "parameters": {"model": "text-embedding-3-small", "options": {}}, "typeVersion": 1}, {"id": "dfd7f734-eb00-4af3-9179-724503422fe4", "name": "Placeholder (File/Content to Upsert)", "type": "n8n-nodes-base.set", "position": [1900, -1000], "parameters": {"mode": "raw", "options": {}, "jsonOutput": "={\n \"Date\": \"{{ $now.format('dd MMM yyyy') }}\",\n \"Time\": \"{{ $now.format('HH:mm ZZZZ z') }}\"\n}\n"}, "typeVersion": 3.4}, {"id": "c54c9458-9b8a-4ef1-a6db-5265729be19d", "name": "Embeddings OpenAI Upserting", "type": "@n8n/n8n-nodes-langchain.embeddingsOpenAi", "position": [2120, -840], "parameters": {"model": "text-embedding-3-small", "options": {}}, "typeVersion": 1}, {"id": "30c18e9e-d047-40d3-8324-f5d0e7892db6", "name": "Insert Documents", "type": "@n8n/n8n-nodes-langchain.vectorStoreSupabase", "position": [2920, -1840], "parameters": {"mode": "insert", "options": {}, "tableName": {"__rl": true, "mode": "list", "value": "Kadampa", "cachedResultName": "Kadampa"}}, "typeVersion": 1}, {"id": "3c0ed0ee-9134-4b4e-bcfd-632dd67a57da", "name": "Retrieve Rows from Table", "type": "n8n-nodes-base.supabase", "position": [3960, -1380], "parameters": {"tableId": "n8n", "operation": "getAll", "returnAll": true}, "typeVersion": 1}, {"id": "53aca1b4-31e8-4699-b158-673623bc9b95", "name": "Sticky Note4", "type": "n8n-nodes-base.stickyNote", "position": [2620, -1160], "parameters": {"color": 6, "width": 1587.0771183771394, "height": 537.3056597675153, "content": "# DELETION\n\nAt the moment n8n does not have a built-in Supabase Node to delete records in a Vector Database. For this you would typically use the HTTP Request node to make an authorized API call to Supabase. \n\n## HTTP Request Node\n\nUse this node to send a DELETE request to your Supabase instance.\n\n- Supabase API Endpoint: Use the appropriate URL for your Supabase project. The endpoint will typically look like this: [https://<your-supabase-ref>.supabase.co/rest/v1/<your-vector-table>](https://supabase.com/docs/guides/api). Replace `<your-supabase-ref>` and `<your-vector-table>` with your details.\n### HEADERS:\n- apikey: Your Supabase API key.\n- Authorization: Bearer token with your Supabase JWT.\n- Query Parameters: Use query parameters to specify which record(s) to delete. For example, `?id=eq.<your-record-id>` where `<your-record-id>` is the specific record ID you want to delete \n(You can also reference back to the **Retrieve Rows From Table** Node to get the ID dynamically)\n\nEnsure you have the necessary permissions set up in Supabase to delete records through the API.\n\nPlease refer to the official n8n documentation for more detailed information on using the [HTTP Request Node](https://docs.n8n.io/integrations/builtin/core-nodes/n8n-nodes-base.httprequest/).\n\n_Note:_ Deleting records is a sensitive operation, so make sure that your permissions are correctly configured and that you are targeting the correct records to avoid unwanted data loss."}, "typeVersion": 1}, {"id": "4ffaccdb-9e0f-464d-9284-7771f6599fd8", "name": "Update Documents", "type": "@n8n/n8n-nodes-langchain.vectorStoreSupabase", "position": [2100, -1000], "parameters": {"id": "1", "mode": "update", "options": {"queryName": "match_documents"}, "tableName": {"__rl": true, "mode": "list", "value": "n8n", "cachedResultName": "n8n"}}, "typeVersion": 1}], "pinData": {}, "connections": {"Google Drive": {"main": [[{"node": "Insert Documents", "type": "main", "index": 0}]]}, "OpenAI Chat Model": {"ai_languageModel": [[{"node": "Question and Answer Chain", "type": "ai_languageModel", "index": 0}]]}, "Retrieve by Query": {"ai_vectorStore": [[{"node": "Vector Store Retriever", "type": "ai_vectorStore", "index": 0}]]}, "Default Data Loader": {"ai_document": [[{"node": "Insert Documents", "type": "ai_document", "index": 0}]]}, "Vector Store Retriever": {"ai_retriever": [[{"node": "Question and Answer Chain", "type": "ai_retriever", "index": 0}]]}, "Question and Answer Chain": {"main": [[{"node": "Customize Response", "type": "main", "index": 0}]]}, "When chat message received": {"main": [[{"node": "Question and Answer Chain", "type": "main", "index": 0}]]}, "Embeddings OpenAI Insertion": {"ai_embedding": [[{"node": "Insert Documents", "type": "ai_embedding", "index": 0}]]}, "Embeddings OpenAI Retrieval": {"ai_embedding": [[{"node": "Retrieve by Query", "type": "ai_embedding", "index": 0}]]}, "Embeddings OpenAI Upserting": {"ai_embedding": [[{"node": "Update Documents", "type": "ai_embedding", "index": 0}]]}, "Recursive Character Text Splitter1": {"ai_textSplitter": [[{"node": "Default Data Loader", "type": "ai_textSplitter", "index": 0}]]}, "Placeholder (File/Content to Upsert)": {"main": [[{"node": "Update Documents", "type": "main", "index": 0}]]}}}How to Import This Workflow
- 1Copy the workflow JSON above using the Copy Workflow JSON button.
- 2Open your n8n instance and go to Workflows.
- 3Click Import from JSON and paste the copied workflow.
Don't have an n8n instance? Start your free trial at n8nautomation.cloud
Related Templates
Ask questions about a PDF using AI
Effortlessly transform your Google Drive PDFs into an interactive knowledge base with this powerful AI workflow. This n8n automation connects your Google Drive files, processes them with OpenAI embeddings, and stores them in a Pinecone vector database, allowing you to ask questions and receive intelligent answers directly from your document content. When a new PDF is uploaded to Google Drive, the workflow automatically extracts its text, splits it into manageable chunks using the Recursive Character Text Splitter, generates embeddings via OpenAI, and then inserts this structured data into Pinecone for efficient retrieval. Later, by clicking the 'Chat' button, you can engage in a natural language conversation with your document, powered by the OpenAI Chat Model and the Question and Answer Chain, which retrieves relevant information from Pinecone. This is ideal for researchers needing to quickly extract insights from large reports, legal professionals analyzing contracts, or businesses creating searchable knowledge bases from their documentation, saving countless hours of manual review and information searching.
Chat with PDF docs using AI (quoting sources)
Chat with PDF docs using AI (quoting sources) Efficiently extract information and generate AI-powered responses directly from your Google Drive PDF documents with this powerful n8n workflow. This automation connects Google Drive, Pinecone, and OpenAI to enable intelligent querying of your document library. When you manually trigger the workflow, it first retrieves a specified PDF from Google Drive using the Download file node. The document content is then processed by the Recursive Character Text Splitter and embedded into a Pinecone vector store using the Embeddings OpenAI and Add to Pinecone vector store nodes, making it searchable. For each query, the Get top chunks matching query node retrieves the most relevant sections from Pinecone, which are then fed to the OpenAI Chat Model via the Answer the query based on chunks node. This allows the AI to provide accurate answers, complete with citations back to the original document sections, thanks to the Structured Output Parser. This workflow is ideal for researchers, legal professionals, and anyone needing to quickly find specific information within large PDF archives, saving significant time and effort in manual document review and ensuring factual accuracy in AI-generated summaries or answers.
Chat with Postgresql Database
Empower your users to interact with your PostgreSQL database using natural language by automating the process of querying and retrieving information. This workflow connects a chat interface, triggered by a new message, to an AI Agent that leverages OpenAI's powerful language model to understand user requests. The AI Agent intelligently utilizes a suite of PostgreSQL tools, including "Get Table Definition," "Execute SQL Query," and "Get DB Schema and Tables List," to dynamically fetch database schema, generate appropriate SQL queries, and execute them against your database. Chat history is maintained using an AI memory buffer, allowing for contextual conversations. This solution is ideal for support teams needing quick data lookups, business analysts exploring data without writing SQL, or developers building interactive data dashboards. It eliminates the need for manual SQL query writing, speeds up data access, and reduces the training burden for non-technical users, saving significant time and resources while improving data accessibility.