Integrate LangChain Chat Trigger in your LLM apps and 422+ apps and services

This n8n workflow demonstrates how to manage your Qdrant vector store when there is a need to keep it in sync with local files. It covers creating, updating and deleting vector store records ensuring our chatbot assistant is never outdated or misleading. Disclaimer This workflow depends on local files accessed through the local filesystem and so will only work on a self-hosted version of n8n at this time. It is possible to amend this workflow to work on n8n cloud by replacing the local file trigger and read file nodes. How it works A local directory where bank statements are downloaded to is monitored via a local file trigger. The trigger watches for the file create, file changed and file deleted events. When a file is created, its contents are uploaded to the vector store. When a file is updated, its previous records are replaced. When the file is deleted, the corresponding records are also removed from the vector store. A simple Question and Answer Chatbot is setup to answer any questions about the bank statements in the system. Requirements A self-hosted version of n8n. Some of the nodes used in this workflow only work with the local filesystem. Qdrant instance to store the records. Customising the workflow This workflow can also work with remote data. Try integrating accounting or CRM software to build a managed system for payroll, invoices and more. Want to go fully local? A version of this workflow is available which uses Ollama instead. You can download this template here: https://drive.google.com/file/d/189F1fNOiw6naNSlSwnyLVEm_Ho_IFfdM/view?usp=sharing

This n8n workflow is a fun way to query and search over your credentials on your n8n instance. Good to know Your credentials should remain safe as this workflow does not decrypt or use any decrypted data. Example Usage "Which workflows are using Slack and Google Calendar?" "Which workflows have AI in their name but are not using openAI?" How it works Using the n8n API, it fetches all workflow data on the instance. Workflow data contains references to credentials used so this will be extracted. With some necessary reformatting, the workflows and their credentials metadata are stored to a SQLite database. Next, an AI agent is used with a custom SQL tool that reads the SQLite database created in the previous step. The AI agent is instructed to perform SQL queries against our workflow credential table when asked about credentials by the user. Requirements You'll need an n8n API key. Please note that only workflows will be scoped to your API key. Customising the workflow Add extra table fields to the SQLite database to answer even more complex queries such as: workflow status to differentiate between active and inactive workflows.

🚀 Local Multi-LLM Testing & Performance Tracker This workflow is perfect for developers, researchers, and data scientists benchmarking multiple LLMs with LM Studio. It dynamically fetches active models, tests prompts, and tracks metrics like word count, readability, and response time, logging results into Google Sheets. Easily adjust temperature 🔥 and top P 🎯 for flexible model testing. Level of Effort: 🟢 Easy – Minimal setup with customizable options. Setup Steps: Install LM Studio and configure models. Update IP to connect to LM Studio. Create a Google Sheet for result tracking. Key Outcomes: Benchmark LLM performance. Automate results in Google Sheets for easy comparison. Version 1.0

Who is this for This workflow is perfect for teams and individuals who manage extensive data in Notion and need a quick, AI-powered way to interact with their databases. If you're looking to streamline your knowledge management, automate searches, and get faster insights from your Notion databases, this workflow is for you. It’s ideal for support teams, project managers, or anyone who needs to query specific data across multiple records or within individual pages of their Notion setup. Check out the Notion template this Assistant is set up to use: https://www.notion.so/templates/knowledge-base-ai-assistant-with-n8n How it works The Notion Database Assistant uses an AI Agent built with Retrieval-Augmented Generation (RAG) to query this Knowledge Base style Notion database. The assistant can search across multiple properties like tags or question and retrieves content from inside individual Notion pages for additional context. Key features include: Querying the database with flexible filters. Searching within individual Notion pages and extracting relevant blocks. Providing a reference link to the exact Notion pages used to inform its responses, ensuring transparency and easy verification. This assistant uses two HTTP request tools—one for querying the Notion database and another for pulling data from within specific pages. It streamlines knowledge retrieval, offering a conversational, AI-driven way to interact with large datasets. Set up Find basic set up instructions inside the workflow itself or watch a quickstart video 👇

This workflow allows you to ask questions about a PDF document. The answers are provided by an AI model of your choice, and the answer includes a citation pointing to the information it used. You can use n8n’s built-in chat interface to ask the questions, or you could customise this workflow to use another one (e.g. Slack, Teams, etc.) Example The workflow is set up with the Bitcoin whitepaper. So you could ask things like: Question: “Which email provider does the creator of Bitcoin use?“ Answer: “GMX [Bitcoin whitepaper.pdf, lines 1-35]” Requirements A Pinecone account (they have a free tier at the time of writing that is easily enough for this workflow) Access to a large language model (e.g. an OpenAI account) Customizing this workflow The workflow only reads in one document, but you could customise it to read in all the documents in a folder (or more). The workflow is set up to use GPT 3.5, but you could swap that out for any other model (including self-hosted ones).

Are you a popular tech startup accelerator (named after a particular higher order function) overwhelmed with 1000s of pitch decks on a daily basis? Wish you could filter through them quickly using AI but the decks are unparseable through conventional means? Then you're in luck! This n8n template uses Multimodal LLMs to parse and extract valuable data from even the most overly designed pitch decks in quick fashion. Not only that, it'll also create the foundations of a RAG chatbot at the end so you or your colleagues can drill down into the details if needed. With this template, you'll scale your capacity to find interesting companies you'd otherwise miss! Requires n8n v1.62.1+ How It Works Airtable is used as the pitch deck database and PDF decks are downloaded from it. An AI Vision model is used to transcribe each page of the pitch deck into markdown. An Information Extractor is used to generate a report from the transcribed markdown and update required information back into pitch deck database. The transcribed markdown is also uploaded to a vector store to build an AI chatbot which can be used to ask questions on the pitch deck. Check out the sample Airtable here: https://airtable.com/appCkqc2jc3MoVqDO/shrS21vGqlnqzzNUc How To Use This template depends on the availability of the Airtable - make a duplicate of the airtable (link) and its columns before running the workflow. When a new pitchdeck is received, enter the company name into the Name column and upload the pdf into the File column. Leave all other columns blank. If you have the Airtable trigger active, the execution should start immediately once the file is uploaded. Otherwise, click the manual test trigger to start the workflow. When manually triggered, all "new" pitch decks will be handled by the workflow as separate executions. Requirements OpenAI for LLM Airtable For Database and Interface Qdrant for Vector Store Customising This Workflow Extend this starter template by adding more AI agents to validate claims made in the pitch deck eg. Linkedin Profiles, Page visits, Reviews etc.

This workflow allows you to ask questions about the data in a Google Sheet over a chat interface. It uses n8n's built-in chat, but could be modified to work with Slack, Teams, WhatsApp, etc. Behind the scenes, the workflow uses GPT4, so you'll need to have an OpenAI API key that supports it. How it works The workflow uses an AI agent with custom tools that call a sub-workflow. That sub-workflow reads the Google Sheet and returns information from it. Because models have a context window (and therefore a maximum number of characters they can accept), we can't pass the whole Google Sheet to GPT - at least not for big sheets. So we provide three ways of querying less data, that can be used in combination to answer questions. Those three functions are: List all the columns in the sheet Get all values of a single column Get all values of a single row Note that to use this template, you need to be on n8n version 1.19.4 or later.

This n8n workflow demonstrates creating a recipe recommendation chatbot using the Qdrant vector store recommendation API. Use this example to build recommendation features in your AI Agents for your users. How it works For our recipes, we'll use HelloFresh's weekly course and recipes for data. We'll scrape the website for this data. Each recipe is split, vectorised and inserted into a Qdrant Collection using Mistral Embeddings Additionally the whole recipe is stored in a SQLite database for later retrieval. Our AI Agent is setup to recommend recipes from our Qdrant vector store. However, instead of the default similarity search, we'll use the Recommendation API instead. Qdrant's Recommendation API allows you to provide a negative prompt; in our case, the user can specify recipes or ingredients to avoid. The AI Agent is now able to suggest a recipe recommendation better suited for the user and increase customer satisfaction. Requirements Qdrant vector store instance to save the recipes Mistral.ai account for embeddings and LLM agent Customising the workflow This workflow can work for a variety of different audiences. Try different sets of data such as clothes, sports shoes, vehicles or even holidays.

Who is this for? This workflow is for everyone who wants to have easier access to their Odoo sales data without complex queries. Use Case To have a clear overview of your sales data in Odoo you typically needs to extract data from it manually to analyse it. This workflow uses OpenAI's language models to create an intelligent chatbot that provides conversational access to your Odoo sales opportunity data. How it works Creates a summary of all Odoo sales opportunities using OpenAI Uses that summary as context for the OpenAI chat model Keeps the summary up to date using a schedule trigger Set up steps: Configure the Odoo credentials Configure OpenAI credentials Toggle "Make Chat Publicly Available" from the Chat Trigger node.

This workflow allows you to ask questions about data stored in a database using AI. To use it, you'll need an OpenAI API key (although you could also swap in a model from another service). Supported databases: Postgres MySQL SQLite The workflow uses n8n's embedded chat, but you could also modify it to work with a chat service such as Slack, MS Teams or WhatsApp. Note that to use this template, you need to be on n8n version 1.19.4 or later.

This is a workflow that tries to answer user queries using the standard GPT-4 model. If it can't answer, it sends a message to Slack to ask for human help. It prompts the user to supply an email address. This workflow is used in Advanced AI examples | Ask a human in the documentation. To use this workflow: Load it into your n8n instance. Add your credentials as prompted by the notes. Configure the Slack node to use your Slack details, or swap out Slack for a different service.

This n8n workflows builds another example of creating a knowledgebase assistant but demonstrates how a more deliberate and targeted approach to ingesting the data can produce much better results for your chatbot. In this example, a government tax code policy document is used. Whilst we could split the document into chunks by content length, we often lose the context of chapters and sections which may be required by the user. Our approach then is to first split the document into chapters and sections before importing into our vector store. Additionally, using metadata correctly is key to allow filtering and scoped queries. Example Human: "Tell me about what the tax code says about cargo for intentional commerce?" AI: "Section 11.25 of the Texas Property Tax Code pertains to "MARINE CARGO CONTAINERS USED EXCLUSIVELY IN INTERNATIONAL COMMERCE." In this section, a person who is a citizen of a foreign country or an en..." How it works The tax code policy document is downloaded as a zip file from the government website and its pages are extracted as separate chapters. Each chapter is then parsed and split into its sections using data manipulation expressions. Each section is then inserted into our Qdrant vector store tagged with its source, chapter and section numbers as metadata. When our AI Agent needs to retrieve data from our vector store, we use a custom workflow tool to perform the query to Qdrant. Because we're relying on Qdrant's advanced filtering capabilities, we perform the search using the Qdrant API rather than the Qdrant node. When the AI Agent, needs to pull full wording or extracts, we can use Qdrant's scroll API and metadata filtering to do so. This makes Qdrant behave like a key-value store for our document. Requirements A Qdrant instance is required for the vector store and specifically for it's filtering functionality. Mistral.ai account for Embeddings and AI models. Customising this workflow Depending on your use-case, consider returning actual PDF pages (or links) to the user for the extra confirmation and to build trust. Not using Mistral? You are able to replace but note to match the distance and dimension size of Qdrant collection to your chosen embedding model.

This is a demo workflow to showcase how to use Supabase to embed a document, retrieve information from the vector store via chat and update the database. Setup steps: set your credentials for Supabase set your credentials for an AI model of your choice set credentials for any service you want to use to upload documents please follow the guidelines in the workflow itself (Sticky Notes) Feedback & Questions If you have any questions or feedback about this workflow - Feel free to get in touch at [email protected]

Chat with local LLMs using n8n and Ollama This n8n workflow allows you to seamlessly interact with your self-hosted Large Language Models (LLMs) through a user-friendly chat interface. By connecting to Ollama, a powerful tool for managing local LLMs, you can send prompts and receive AI-generated responses directly within n8n. Use cases Private AI Interactions Ideal for scenarios where data privacy and confidentiality are important. Cost-Effective LLM Usage Avoid ongoing cloud API costs by running models on your own hardware. Experimentation & Learning A great way to explore and experiment with different LLMs in a local, controlled environment. Prototyping & Development Build and test AI-powered applications without relying on external services. How it works When chat message received: Captures the user's input from the chat interface. Chat LLM Chain: Sends the input to the Ollama server and receives the AI-generated response. Delivers the LLM's response back to the chat interface. Set up steps Make sure Ollama is installed and running on your machine before executing this workflow. Edit the Ollama address if different from the default.

OpenAI Assistant is a powerful tool, but at the time of writing it doesn't automatically remember past messages from a conversation. This workflow demonstrates how to get around this, by managing the chat history in n8n and passing it to the assistant when required. This makes it possible to use OpenAI Assistant for chatbot use cases. Note that to use this template, you need to be on n8n version 1.28.0 or later.

Create a recommendation tool without hallucinations based on RAG with the Qdrant Vector database. This example is based on movie recommendations on the IMDB-top1000 dataset. You can provide your wishes and your "big no's" to the chatbot, for example: "A movie about wizards but not Harry Potter", and get top-3 recommendations. How it works a video with the full design process Upload IMDB-1000 dataset to Qdrant Vector Store, embedding movie descriptions with OpenAI; Set up an AI agent with a chat. This agent will call a workflow tool to get movie recommendations based on a request written in the chat; Create a workflow which calls Qdrant's Recommendation API to retrieve top-3 recommendations of movies based on your positive and negative examples. Set Up Steps You'll need to create a free tier Qdrant Cluster (Qdrant can also be used locally; it's open-sourced) and set up API credentials You'll OpenAI credentials You'll need GitHub credentials & to upload the IMDB Kaggle dataset to your GitHub.

This workflow was presented at the AI Developers meet up in San Fransico on 24 July, 2024. AI workflows Categorize incoming Gmail emails and assign custom Gmail labels. This example uses the Text Classifier node, simplifying this usecase. Ingest a PDF into a Pinecone vector store and chat with it (RAG example) AI Agent example showcasing the HTTP Request tool. We teach the agent how to check availability on a Google Calendar and book an appointment.

Overview This workflow aims at providing data visualization to a native SQL Agent. Together, they can help with fostering data analysis and data visualization within a team. It uses the native SQL Agent that works well and adds some visualization capabilities thanks to OpenAI Structured Output and Quickchart.io. How it works The first part of the workflow is a regular SQL Agent: it connects to a Database and is able to query it and translate the response in a human format. Then, the Text Classifier is deciding if the user would benefit from a chart, supporting the SQL Agent's response. If it does, then it executes the subworkflow to dynamically generate a chart and append the chart to the response from the SQL Agent. If it doesn't, then the SQL Agent response is directly outputted. The sub-workflow calls OpenAI through the HTTP Request node to retrieve a chart definition. In the "set response" node, the chart definition is added at the end of a quickchart.io URL - the URL to the chart image. It is sent back to the AI Agent. How to use it Use an existing or create a new database. For example, I've used this Kaggle dataset and uploaded it to a Supabase DB. Add the PostgreSQL or MySQL credentials. Alternatively, you can use SQLite binary files (check this template). Activate the workflow. Start chatting with the AI SQL Agent. If the Text Classifier considers a chart would be useful, it will generate a chart in addition to the response from the SQL Agent. Notes The full Quickchart.io specifications have not been integrated, thus there are some possible glitches (e.g., due to the size of the graph, radar graphs are not displayed properly).

Who is this for? This workflow is designed for: Database administrators and developers working with MongoDB Content managers handling movie databases Organizations looking to implement AI-powered search and recommendation systems Developers interested in combining LangChain, OpenAI, and MongoDB capabilities What problem does this workflow solve? Traditional database queries can be complex and require specific MongoDB syntax knowledge. This workflow addresses: The complexity of writing MongoDB aggregation pipelines The need for natural language interaction with movie databases The challenge of maintaining user preferences and favorites The gap between AI language models and database operations What this workflow does This workflow creates an intelligent agent that: Accepts natural language queries about movies Translates user requests into MongoDB aggregation pipelines Queries a movie database containing detailed information including: Plot summaries Genre classifications Cast and director information Runtime and release dates Ratings and awards Provides contextual responses using OpenAI's language model Allows users to save favorite movies to the database Maintains conversation context using a window buffer memory Setup Required Credentials: OpenAI API credentials MongoDB connection details Node Configuration: Configure the MongoDB connection in the MongoDBAggregate node Set up the OpenAI Chat Model with your API key Ensure the webhook trigger is properly configured for receiving chat messages Database Requirements: A MongoDB collection named "movies" with the specified document structure Proper indexes for efficient querying Appropriate user permissions for read/write operations How to customize this workflow Modify the Document Structure: Update the tool description in the MongoDBAggregate node to match your collection schema Adjust the aggregation pipeline templates for your specific use case Enhance the AI Agent: Customize the prompt in the "AI Agent - Movie Recommendation" node Modify the window buffer memory size based on your context needs Add additional tools for more functionality Extend Functionality: Add more MongoDB operations beyond aggregation Implement additional workflows for different types of queries Create custom error handling and validation Add user authentication and rate limiting Integration Options: Connect to external APIs for additional movie data Add webhook endpoints for different platforms Implement caching mechanisms for frequent queries Add data transformation nodes for specific output formats This workflow serves as a foundation that can be adapted to various use cases beyond movie recommendations, such as e-commerce product search, content management systems, or any scenario requiring intelligent database interaction.

This n8n workflow demonstrates how to create an agent using LangChain and SQLite. The agent can understand natural language queries and interact with a SQLite database to provide accurate answers. 💪 🚀 Setup Run the top part of the workflow once. It downloads the example SQLite database, extracts from a ZIP file and saves locally (chinook.db). 🗣️ Chatting with Your Data Send a message in a chat window. Locally saved SQLite database loads automatically. User's chat input is combined with the binary data. The LangChain Agend node gets both data and begins to work. The AI Agent will process the user's message, perform necessary SQL queries, and generate a response based on the database information. 🗄️ 🌟 Example Queries Try these sample queries to see the AI Agent in action: "Please describe the database" - Get a high-level overview of the database structure, only one or two queries are needed. "What are the revenues by genre?" - Retrieve revenue information grouped by genre, LangChain agent iterates several time before producing the answer. The AI Agent will store the final answer in its memory, allowing for context-aware conversations. 💬 Read the full article: 👉 https://blog.n8n.io/ai-agents/

This is the first version of a template for a RAG/GenAI App using WordPress content. As creating, sharing, and improving templates brings me joy 😄, feel free to reach out on LinkedIn if you have any ideas to enhance this template! How It Works This template includes three workflows: Workflow 1**: Generate embeddings for your WordPress posts and pages, then store them in the Supabase vector store. Workflow 2**: Handle upserts for WordPress content when edits are made. Workflow 3**: Enable chat functionality by performing Retrieval-Augmented Generation (RAG) on the embedded documents. Why use this template? This template can be applied to various use cases: Build a GenAI application that requires embedded documents from your website's content. Embed or create a chatbot page on your website to enhance user experience as visitors search for information. Gain insights into the types of questions visitors are asking on your website. Simplify content management by asking the AI for related content ideas or checking if similar content already exists. Useful for internal linking. Prerequisites Access to Supabase for storing embeddings. Basic knowledge of Postgres and pgvector. A WordPress website with content to be embedded. An OpenAI API key Ensure that your n8n workflow, Supabase instance, and WordPress website are set to the same timezone (or use GMT) for consistency. Workflow 1 : Initial Embedding This workflow retrieves your WordPress pages and posts, generates embeddings from the content, and stores them in Supabase using pgvector. Step 0 : Create Supabase tables Nodes : Postgres - Create Documents Table: This table is structured to support OpenAI embedding models with 1536 dimensions Postgres - Create Workflow Execution History Table These two nodes create tables in Supabase: The documents table, which stores embeddings of your website content. The n8n_website_embedding_histories table, which logs workflow executions for efficient management of upserts. This table tracks the workflow execution ID and execution timestamp. Step 1 : Retrieve and Merge WordPress Pages and Posts Nodes : WordPress - Get All Posts WordPress - Get All Pages Merge WordPress Posts and Pages These three nodes retrieve all content and metadata from your posts and pages and merge them. Important: ** **Apply filters to avoid generating embeddings for all site content. Step 2 : Set Fields, Apply Filter, and Transform HTML to Markdown Nodes : Set Fields Filter - Only Published & Unprotected Content HTML to Markdown These three nodes prepare the content for embedding by: Setting up the necessary fields for content embeddings and document metadata. Filtering to include only published and unprotected content (protected=false), ensuring private or unpublished content is excluded from your GenAI application. Converting HTML to Markdown, which enhances performance and relevance in Retrieval-Augmented Generation (RAG) by optimizing document embeddings. Step 3: Generate Embeddings, Store Documents in Supabase, and Log Workflow Execution Nodes: Supabase Vector Store Sub-nodes: Embeddings OpenAI Default Data Loader Token Splitter Aggregate Supabase - Store Workflow Execution This step involves generating embeddings for the content and storing it in Supabase, followed by logging the workflow execution details. Generate Embeddings: The Embeddings OpenAI node generates vector embeddings for the content. Load Data: The Default Data Loader prepares the content for embedding storage. The metadata stored includes the content title, publication date, modification date, URL, and ID, which is essential for managing upserts. ⚠️ Important Note : Be cautious not to store any sensitive information in metadata fields, as this information will be accessible to the AI and may appear in user-facing answers. Token Management: The Token Splitter ensures that content is segmented into manageable sizes to comply with token limits. Aggregate: Ensure the last node is run only for 1 item. Store Execution Details: The Supabase - Store Workflow Execution node saves the workflow execution ID and timestamp, enabling tracking of when each content update was processed. This setup ensures that content embeddings are stored in Supabase for use in downstream applications, while workflow execution details are logged for consistency and version tracking. This workflow should be executed only once for the initial embedding. Workflow 2, described below, will handle all future upserts, ensuring that new or updated content is embedded as needed. Workflow 2: Handle document upserts Content on a website follows a lifecycle—it may be updated, new content might be added, or, at times, content may be deleted. In this first version of the template, the upsert workflow manages: Newly added content** Updated content** Step 1: Retrieve WordPress Content with Regular CRON Nodes: CRON - Every 30 Seconds Postgres - Get Last Workflow Execution WordPress - Get Posts Modified After Last Workflow Execution WordPress - Get Pages Modified After Last Workflow Execution Merge Retrieved WordPress Posts and Pages A CRON job (set to run every 30 seconds in this template, but you can adjust it as needed) initiates the workflow. A Postgres SQL query on the n8n_website_embedding_histories table retrieves the timestamp of the latest workflow execution. Next, the HTTP nodes use the WordPress API (update the example URL in the template with your own website’s URL and add your WordPress credentials) to request all posts and pages modified after the last workflow execution date. This process captures both newly added and recently updated content. The retrieved content is then merged for further processing. Step 2 : Set fields, use filter Nodes : Set fields2 Filter - Only published and unprotected content The same that Step 2 in Workflow 1, except that HTML To Makrdown is used in further Step. Step 3: Loop Over Items to Identify and Route Updated vs. Newly Added Content Here, I initially aimed to use 'update documents' instead of the delete + insert approach, but encountered challenges, especially with updating both content and metadata columns together. Any help or suggestions are welcome! :) Nodes: Loop Over Items Postgres - Filter on Existing Documents Switch Route existing_documents (if documents with matching IDs are found in metadata): Supabase - Delete Row if Document Exists: Removes any existing entry for the document, preparing for an update. Aggregate2: Used to aggregate documents on Supabase with ID to ensure that Set Fields3 is executed only once for each WordPress content to avoid duplicate execution. Set Fields3: Sets fields required for embedding updates. Route new_documents (if no matching documents are found with IDs in metadata): Set Fields4: Configures fields for embedding newly added content. In this step, a loop processes each item, directing it based on whether the document already exists. The Aggregate2 node acts as a control to ensure Set Fields3 runs only once per WordPress content, effectively avoiding duplicate execution and optimizing the update process. Step 4 : HTML to Markdown, Supabase Vector Store, Update Workflow Execution Table The HTML to Markdown node mirrors Workflow 1 - Step 2. Refer to that section for a detailed explanation on how HTML content is converted to Markdown for improved embedding performance and relevance. Following this, the content is stored in the Supabase vector store to manage embeddings efficiently. Lastly, the workflow execution table is updated. These nodes mirros the **Workflow 1 - Step 3 nodes. Workflow 3 : An example of GenAI App with Wordpress Content : Chatbot to be embed on your website Step 1: Retrieve Supabase Documents, Aggregate, and Set Fields After a Chat Input Nodes: When Chat Message Received Supabase - Retrieve Documents from Chat Input Embeddings OpenAI1 Aggregate Documents Set Fields When a user sends a message to the chat, the prompt (user question) is sent to the Supabase vector store retriever. The RPC function match_documents (created in Workflow 1 - Step 0) retrieves documents relevant to the user’s question, enabling a more accurate and relevant response. In this step: The Supabase vector store retriever fetches documents that match the user’s question, including metadata. The Aggregate Documents node consolidates the retrieved data. Finally, Set Fields organizes the data to create a more readable input for the AI agent. Directly using the AI agent without these nodes would prevent metadata from being sent to the language model (LLM), but metadata is essential for enhancing the context and accuracy of the AI’s response. By including metadata, the AI’s answers can reference relevant document details, making the interaction more informative. Step 2: Call AI Agent, Respond to User, and Store Chat Conversation History Nodes: AI Agent** Sub-nodes: OpenAI Chat Model Postgres Chat Memories Respond to Webhook** This step involves calling the AI agent to generate an answer, responding to the user, and storing the conversation history. The model used is gpt4-o-mini, chosen for its cost-efficiency.

Use n8n to bring data from any API to your AI. This workflow uses the Chat Trigger to provide the chat interface, and the Custom n8n Workflow Tool to call a second workflow that calls the API. The second workflow uses AI functionality to refine the API request based on the user's query. It then makes an API call, and returns the response to the main workflow. This workflow is used in Advanced AI examples | Call an API to fetch data in the documentation. To use this workflow: Load it into your n8n instance. Add your credentials as prompted by the notes. Requires n8n 1.28.0 or above

Purpose Use a lightweight Voice Interface, for you and your entire organization, to interact with an AI Supervisor, a personal AI Assistant, which has access to your custom workflows. You can also connect the supervisor to your already existing Agents. Demo & Explanation How it works After recording a message in the Vagent App, it gets transcribed and sent in combination with a session ID to the registered webhook The Main Agent acts as a router. I interprets the message while using the stored chat history (bound to the session ID) and chooses which tool to use to perform the required action and. Tools on this level are workflows, which contain subordinated Agents. Since the Main Agent interprets the original message, the raw input is passed to the Tools/Sub-Agents as a separate parameter Within the Sub-Agents the actual processing takes place. Each of those has it’s separate chat memory (with a suffix to the main session ID), to achieve a clear separation of concerns Depending on the required action an HTTP Request Tool is called. The result is being formatted in Markdown and returned to the Main Agent with an additional short prompt, so it does not get interpreted by the Main Agent. Drafts are separated from a short message by added indentation (angle brackets). If some information is missing, no tool is called just yet, instead a message is returned back to the user The Main Agent then outputs the result from the called Sub-Agent. If a draft is included, it gets separated from the spoken output Finally the formatted output is returned as response to the webhook. The message is split into a spoken and a text version, which enables the App to read out loud unnecessary information like drafts in this example See the full documentation of Vagent: https://vagent.io/docs Setup Import this workflow into your n8n instance Follow the instructions given in the sticky notes on the canvas Setup your credentials. OpenAI can be replaced by another LLM in the workflow, but is required for the App to work. Google Calendar and Notion are required for all scenarios to work Copy the Webhook URL from the Webhook node of the main workflow Download the Vagent App from https://vagent.io In the settings paste your OpenAI API Token, the Webhook URL and the password defined for Header Auth Now you can use the App to interact with the Multi-Agent using your Voice by tapping the Mic symbol in the App to record your message. To use the chat trigger (for testing) properly, temporarily disable the nodes after the Tools Agent.