LLM/n8n-install
1
0
Fork 0
mirror of https://github.com/kossakovsky/n8n-install.git synced 2026-03-07 22:33:11 +00:00
Code Issues Packages Projects Releases Wiki Activity

Update V3_Local_Agentic_RAG_AI_Agent.json

Browse Source
This commit is contained in:
Cole Medin
2025-03-05 21:41:14 -06:00
parent b303716c89
commit af22927da2
1 changed files with 6 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
n8n/backup/workflows/V3_Local_Agentic_RAG_AI_Agent.json
View File

@@ -272,7 +272,7 @@
"promptType": "define",
"text": "={{ $json.chatInput }}",
"options": {
"systemMessage": "You are a personal assistant who helps answer questions from a corpus of documents. The documents are either text based (Txt, docs, extracted PDFs, etc.) or tabular data (CSVs or Excel documents).\n\nYou are given tools to perform RAG in the 'documents' table, look up the documents available in your knowledge base in the 'document_metadata' table, extract all the text from a given document, and query the tabular files with SQL in the 'document_rows' table.\n\nAlways start by performing RAG unless the question requires a SQL query for tabular data (fetching a sum, finding a max, something a RAG lookup would be unreliable for). If RAG doesn't help, then look at the documents that are available to you, find a few that you think would contain the answer, and then analyze those.\n\nAlways tell the user if you didn't find the answer. Don't make something up just to please them."
"systemMessage": "You are a personal assistant who helps answer questions from a corpus of documents. The documents are either text based (Txt, docs, extracted PDFs, etc.) or tabular data (CSVs or Excel documents).\n\nYou are given tools to perform RAG in the 'documents' table, look up the documents available in your knowledge base in the 'document_metadata' table, extract all the text from a given document, and query the tabular files with SQL in the 'document_rows' table.\n\nAlways start by performing RAG unless the users asks you to check a document or the question requires a SQL query for tabular data (fetching a sum, finding a max, something a RAG lookup would be unreliable for). If RAG doesn't help, then look at the documents that are available to you, find a few that you think would contain the answer, and then analyze those.\n\nAlways tell the user if you didn't find the answer. Don't make something up just to please them."
}
},
"id": "8d6daa23-a646-4592-a449-02cf50e2c953",
@@ -546,7 +546,7 @@
"descriptionType": "manual",
"toolDescription": "Given a file ID, fetches the text from the document.",
"operation": "executeQuery",
"query": "SELECT \n string_agg(content, ' ') as document_text\nFROM documents\n WHERE metadata->>'file_id' = $1\nGROUP BY metadata->>'file_id';",
"query": "SELECT \n string_agg(content, ' ') as document_text\nFROM documents_pg\n WHERE metadata->>'file_id' = $1\nGROUP BY metadata->>'file_id';",
"options": {
"queryReplacement": "={{ $fromAI('file_id') }}"
}
@@ -868,7 +868,7 @@
},
{
"parameters": {
"content": "## 🚀 n8n Local AI Agentic RAG Template\n\n**Author:** [Cole Medin](https://www.youtube.com/@ColeMedin)\n\n## What is this?\nThis template provides an entirely local implementation of an **Agentic RAG (Retrieval Augmented Generation)** system in n8n that can be extended easily for your specific use case and knowledge base. Unlike standard RAG which only performs simple lookups, this agent can reason about your knowledge base, self-improve retrieval, and dynamically switch between different tools based on the specific question. This workflow can be used with the [local AI package](https://github.com/coleam00/local-ai-packaged) for your LLMs (Ollama) and database (Supabase).\n\n## Why Agentic RAG?\nStandard RAG has significant limitations:\n- Poor analysis of numerical/tabular data\n- Missing context due to document chunking\n- Inability to connect information across documents\n- No dynamic tool selection based on question type\n\n## What makes this template powerful:\n- **Intelligent tool selection**: Switches between RAG lookups, SQL queries, or full document retrieval based on the question\n- **Complete document context**: Accesses entire documents when needed instead of just chunks\n- **Accurate numerical analysis**: Uses SQL for precise calculations on spreadsheet/tabular data\n- **Cross-document insights**: Connects information across your entire knowledge base\n- **Multi-file processing**: Handles multiple documents in a single workflow loop\n- **Efficient storage**: Uses JSONB in Supabase to store tabular data without creating new tables for each CSV\n\n## Getting Started\n1. Run the table creation nodes first to set up your database tables in Supabase\n2. Upload your documents to the folder on your computer that is mounted to /data/shared in the n8n container. This folder by default is the \"shared\" folder in the local AI package.\n3. The agent will process them automatically (chunking text, storing tabular data in Supabase)\n4. Start asking questions that leverage the agent's multiple reasoning approaches\n\n## Customization\nThis template provides a solid foundation that you can extend by:\n- Tuning the system prompt for your specific use case\n- Adding document metadata like summaries\n- Implementing more advanced RAG techniques\n- Optimizing for larger knowledge bases\n\n---\n\nThe non-local (\"cloud\") version of this Agentic RAG agent can be [found here](https://github.com/coleam00/ottomator-agents/tree/main/n8n-agentic-rag-agent).",
"content": "## 🚀 n8n Local AI Agentic RAG Template\n\n**Author:** [Cole Medin](https://www.youtube.com/@ColeMedin)\n\n## What is this?\nThis template provides an entirely local implementation of an **Agentic RAG (Retrieval Augmented Generation)** system in n8n that can be extended easily for your specific use case and knowledge base. Unlike standard RAG which only performs simple lookups, this agent can reason about your knowledge base, self-improve retrieval, and dynamically switch between different tools based on the specific question. This workflow can be used with the [local AI package](https://github.com/coleam00/local-ai-packaged) for your LLMs (Ollama) and database (Supabase or \"vanilla\" Postgres).\n\n## Why Agentic RAG?\nStandard RAG has significant limitations:\n- Poor analysis of numerical/tabular data\n- Missing context due to document chunking\n- Inability to connect information across documents\n- No dynamic tool selection based on question type\n\n## What makes this template powerful:\n- **Intelligent tool selection**: Switches between RAG lookups, SQL queries, or full document retrieval based on the question\n- **Complete document context**: Accesses entire documents when needed instead of just chunks\n- **Accurate numerical analysis**: Uses SQL for precise calculations on spreadsheet/tabular data\n- **Cross-document insights**: Connects information across your entire knowledge base\n- **Multi-file processing**: Handles multiple documents in a single workflow loop\n- **Efficient storage**: Uses JSONB in Supabase to store tabular data without creating new tables for each CSV\n\n## Getting Started\n1. Run the table creation nodes first to set up your database tables in Supabase\n2. Upload your documents to the folder on your computer that is mounted to /data/shared in the n8n container. This folder by default is the \"shared\" folder in the local AI package.\n3. The agent will process them automatically (chunking text, storing tabular data in Supabase)\n4. Start asking questions that leverage the agent's multiple reasoning approaches\n\n## Customization\nThis template provides a solid foundation that you can extend by:\n- Tuning the system prompt for your specific use case\n- Adding document metadata like summaries\n- Implementing more advanced RAG techniques\n- Optimizing for larger knowledge bases\n\n---\n\nThe non-local (\"cloud\") version of this Agentic RAG agent can be [found here](https://github.com/coleam00/ottomator-agents/tree/main/n8n-agentic-rag-agent).",
"height": 1320,
"width": 540,
"color": 6
@@ -976,9 +976,9 @@
"parameters": {
"model": {
"__rl": true,
"value": "qwen2.5:7b-8k",
"value": "qwen2.5:14b-8k",
"mode": "list",
"cachedResultName": "qwen2.5:7b-8k"
"cachedResultName": "qwen2.5:14b-8k"
},
"options": {}
},
@@ -1483,7 +1483,7 @@
"settings": {
"executionOrder": "v1"
},
"versionId": "0bab9394-1560-49a0-a8c3-1ec43a250b1b",
"versionId": "cd039b65-853f-4cde-892c-0ad66da0ef02",
"meta": {
"templateCredsSetupCompleted": true,
"instanceId": "73cb7a3e883df514bb47e8d1b34526d30e2abb8f56cd99f10d5948a1e11b25aa"