Jimleuk
Founder @ Subworkflow.ai - the fastest way to build durable RAG applications. Freelance AI Automation Engineer based in London, UK. Since 2024, my n8n templates have documented my journey into applied AI and have helped hundreds of businesses and organisations get up to speed with AI automation. Today, I continue to explore use-cases as AI evolves and occasionally upload templates which I find novel and interesting. Subscribe to the RSS Feed: https://cdn.subworkflow.ai/n8n-templates/rss.xml
Categories
Templates by Jimleuk
Building your first WhatsApp chatbot
This n8n template builds a simple WhatsApp chabot acting as a Sales Agent. The Agent is backed by a product catalog vector store to better answer user's questions. This template is intended to help introduce n8n users interested in building with WhatsApp. How it works This template is in 2 parts: creating the product catalog vector store and building the WhatsApp AI chatbot. A product brochure is imported via HTTP request node and its text contents extracted. The text contents are then uploaded to the in-memory vector store to build a knowledgebase for the chatbot. A WhatsApp trigger is used to capture messages from customers where non-text messages are filtered out. The customer's message is sent to the AI Agent which queries the product catalogue using the vector store tool. The Agent's response is sent back to the user via the WhatsApp node. How to use Once you've setup and configured your WhatsApp account and credentials First, populate the vector store by clicking the "Test Workflow" button. Next, activate the workflow to enable the WhatsApp chatbot. Message your designated WhatsApp number and you should receive a message from the AI sales agent. Tweak datasource and behaviour as required. Requirements WhatsApp Business Account OpenAI for LLM Customising this workflow Upgrade the vector store to Qdrant for persistance and production use-cases. Handle different WhatsApp message types for a more rich and engaging experience for customers.
By JimleukRespond to WhatsApp messages with AI like a pro!
This n8n template demonstrates the beginnings of building your own n8n-powered WhatsApp chatbot! Under the hood, utilise n8n's powerful AI features to handle different message types and use an AI agent to respond to the user. A powerful tool for any use-case! How it works Incoming WhatsApp Trigger provides a way to get messages into the workflow. The message received is extracted and sent through 1 of 4 branches for processing. Each processing branch uses AI to analyse, summarize or transcribe the message so that the AI agent can understand it. The supported types are text, image, audio (voice notes) and video. The AI Agent is used to generate a response generally and uses a wikipedia tool for more complex queries. Finally, the response message is sent back to the WhatsApp user using the WhatsApp node. How to use Once you have setup and configured your WhatsApp account, you'll need to activate your workflow to start processing messages. Good to know: Large media files may negatively impact workflow performance. Requirements WhatsApp Buisness account Google Gemini for LLM. Gemini is used specifically because it can accept audio and video files whereas at time of writing, many other providers like OpenAI's GPT, do not. Customising this workflow For performance reasons, consider detecting large audio and video before sending to the LLM. Pre-processing such files may allow your agent to perform better. Go beyond and create rich and engagement customer experiences by responding using images, audio and video instead of just text!
By JimleukBuild your own N8N workflows MCP server
This n8n template shows you how to create an MCP server out of your existing n8n workflows. With this, any MCP client connected can get more done with powerful end-to-end workflows rather than just simple tools. Designing agent tools for outcome rather than utility has been a long recommended practice of mine and it applies well when it comes to building MCP servers; In gist, agents to be making the least amount of calls possible to complete a task. This is why n8n can be a great fit for MCP servers! This template connects your agent/MCP client (like Claude Desktop) to your existing workflows by allowing the AI to discover, manage and run these workflows indirectly. How it works An MCP trigger is used and attaches 4 custom workflow tools to discover and manage existing workflows to use and 1 custom workflow tool to execute them. We'll introduce an idea of "available" workflows which the agent is allowed to use. This will help limit and avoid some issues when trying to use every workflow such as clashes or non-production. The n8n node is a core node which taps into your n8n instance API and is able to retrieve all workflows or filter by tag. For our example, we've tagged the workflows we want to use with "mcp" and these are exposed through the tool "search workflows". Redis is used as our main memory for keeping track of which workflows are "available". The tools we have are "add Workflow", "remove workflow" and "list workflows". The agent should be able to manage this autonomously. Our approach to allow the agent to execute workflows is to use the Subworkflow trigger. The tricky part is figuring out the input schema for each but was eventually solved by pulling this information out of the workflow's template JSON and adding it as part of the "available" workflow's description. To pass parameters through the Subworkflow trigger, we can do so via the passthrough method - which is that incoming data is used when parameters are not explicitly set within the node. When running, the agent will not see the "available" workflows immediately but will need to discover them via "list" and "search". The human will need to make the agent aware that these workflows will be preferred when answering queries or completing tasks. How to use First, decide which workflows will be made visible to the MCP server. This example uses the tag of "mcp" but you can all workflows or filter in other ways. Next, ensure these workflows have Subworkflow triggers with input schema set. This is how the MCP server will run them. Set the MCP server to "active" which turns on production mode and makes available to production URL. Use this production URL in your MCP client. For Claude Desktop, see the instructions here - https://docs.n8n.io/integrations/builtin/core-nodes/n8n-nodes-langchain.mcptrigger/integrating-with-claude-desktop. There is a small learning curve which will shape how you communicate with this MCP server so be patient and test. The MCP server will work better if there is a focused goal in mind ie. Research and report, rather than just a collection of unrelated tools. Requirements N8N API key to filter for selected workflows. N8N workflows with Subworkflow triggers! Redis for memory and tracking the "available" workflows. MCP Client or Agent for usage such as Claude Desktop - https://claude.ai/download Customising this workflow If your targeted workflows do not use the subworkflow trigger, it is possible to amend the executeTool to use HTTP requests for webhooks. Managing available workflows helps if you have many workflows where some may be too similar for the agent. If this isn't a problem for you however, feel free to remove the concept of "available" and let the agent discover and use all workflows!
By JimleukHost your own AI deep research agent with n8n, Apify and OpenAI o3
This template attempts to replicate OpenAI's DeepResearch feature which, at time of writing, is only available to their pro subscribers. > An agent that uses reasoning to synthesize large amount of online information and complete multi-step research tasks for you. Source Though the inner workings of DeepResearch have not been made public, it is presumed the feature relies on the ability to deep search the web, scrape web content and invoking reasoning models to generate reports. All of which n8n is really good at! Using this workflow, n8n users can enjoy a variation of the Deep Research experience for themselves and their teams at a fraction of the cost. Better yet, learn and customise this Deep Research template for their businesses and/or organisations. Check out the generated reports here: https://jimleuk.notion.site/19486dd60c0c80da9cb7eb1468ea9afd?v=19486dd60c0c805c8e0c000ce8c87acf How it works A form is used to first capture the user's research query and how deep they'd like the researcher to go. Once submitted, a blank Notion page is created which will later hold the final report and the researcher gets to work. The user's query goes through a recursive series of web serches and web scraping to collect data on the research topic to generate partial learnings. Once complete, all learnings are combined and given to a reasoning LLM to generate the final report. The report is then written to the placeholder Notion page created earlier. How to use Duplicate this Notion database template and make sure all Notion related nodes point to it. Sign-up for APIFY.com API Key for web search and scraping services. Ensure you have access to OpenAI's o3-mini model. Alternatively, switch this out for o1 series. You must publish this workflow and ensure the form url is publically accessible. On depth & breadth configuration For more detailed reports, increase depth and breadth but be warned the workflow will take exponentially longer and cost more to complete. The recommended defaults are usually good enough. Depth=1 & Breadth=2 - will take about 5 - 10mins. Depth=1 & Breadth=3 - will take about 15 - 20mins. Dpeth=3 & Breadth=5 - will take about 2+ hours! Customising this workflow I deliberately chose not to use AI-powered scrapers like Firecrawl as I felt these were quite costly and quotas would be quickly exhausted. However, feel free to switch web search and scraping services which suit your environment. Maybe you don't decide to source the web and instead, data collection comes from internal documents instead. This template gives you freedom to change this. Experiment with different Reasoning/Thinking models such as Deepseek and Google's Gemini 2.0. Finally, the LLM prompts could definitely be improved. Refine them to fit your use-case. Credits This template is largely based off the work by David Zhang (dzhng) and his open source implementation of Deep Research: https://github.com/dzhng/deep-research
By JimleukCommunity insights using Qdrant, Python and Information Extractor
This n8n template is one of a 3-part series exploring use-cases for clustering vector embeddings: Survey Insights Customer Insights Community Insights This template demonstrates the Community Insights scenario where HN commments can be quickly grouped by similarity and an AI agent can generate insights on those groupings. With this workflow, Researchers or HN users can quickly breakdown community consensus on a particular topic and identify frequently mentioned positives and negatives. Sample Output: https://docs.google.com/spreadsheets/d/e/2PACX-1vQXaQU9XxsxnUIIeqmmf1PuYRuYtwviVXTv6Mz9Vo6_a4ty-XaJHSeZsptjWXS3wGGDG8Z4u16rvE7l/pubhtml How it works HN comments are imported via the Hacknews API node. Comments are then inserted into a Qdrant collection carefully tagged with the Hackernews API metadata. Comments are then fetched and are put through a clustering algorithm using the Python Code node. The Qdrant points are returned in clustered groups. Each group is looped to fetch the payloads of the points and feed them to the AI agent to summarise and generate insights for. The resulting insights and raw responses are then saved to the Google Spreadsheet for further analysis by the researcher or the HN user. Requirements Works best with lots of comments! Qdrant Vectorstore for storing embeddings. OpenAI account for embeddings and LLM. Customising the Template Adjust clustering parameters which make sense for your data. Adjust sentimentality setting if comments are overwhelmingly negative at times.
By JimleukActioning your meeting next steps using transcripts and AI
This n8n workflow demonstrates how you can summarise and automate post-meeting actions from video transcripts fed into an AI Agent. Save time between meetings by allowing AI handle the chores of organising follow-up meetings and invites. How it works This workflow scans for the calendar for client or team meetings which were held online. Attempts will be made to fetch any recorded transcripts which are then sent to the AI agent. The AI agent summarises and identifies if any follow-on meetings are required. If found, the Agent will use its Calendar Tool to to create the event for the time, date and place for the next meeting as well as add known attendees. Requirements Google Calendar and the ability to fetch Meeting Transcripts (There is a special OAuth permission for this action!) OpenAI account for access to the LLM. Customising the workflow This example only books follow-on meetings but could be extended to generate reports or send emails.
By JimleukBreakdown documents into study notes using templating MistralAI and Qdrant
This n8n workflow takes in a document such as a research paper, marketing or sales deck or company filings, and breaks them down into 3 templates: study guide, briefing doc and timeline. These templates are designed to help a student, associate or clerk quickly summarise, learn and understand the contents to be more productive. Study guide - a short quiz of questions and answered generated by the AI Agent using the contents of the document. Briefing Doc - key information and insights are extracted by the AI into a digestable form. Timeline - key events, durations and people are identified and listed into a simple to understand timeline by the AI How it works A local file trigger watches a local network directory for new documents. New documents are imported into the workflow, its contents extracted and vectorised into a Qdrant vector store to build a mini-knowledgebase. The document then passes through a series of template generating prompts where the AI will perform "research" on the knowledgebase to generate the template contents. Generated study guide, briefing and timeline documents are exported to a designated folder for the user. Requirements Self-hosted version of n8n. Qdrant instance for knowledgebase. Mistral.ai account for embeddings and AI model. Customising your workflow Try adding your own templates or adjusting the existing templates to suit your unique use-case. Anything is quite possible and limited only by your imagination! 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/1VV5R2nW-IhVcFP_k8uEks4LsLRZrHSNG/view?usp=sharing
By JimleukHandling appointment leads and follow-up with Twilio, Cal.com and AI
This n8n workflow builds an appointment scheduling AI agent which can Take enquiries from prospective customers and help them book an appointment by checking appointment availability Where no appointment is booked, the Agent is able to send follow-up messages to re-engage leads. After an appointment is booked, the agent is able reschedule or even cancel the booking for the user without human intervention. For small outfits, this workflow could contribute the necessary "man-power" required to increase business sales. The sample Airtable can be found here: https://airtable.com/appO2nHiT9XPuGrjN/shroSFT2yjf87XAox 2024-10-22 Updated to Cal.com API v2. How it works The customer sends an enquiry via SMS to trigger our workflow. For this trigger, we'll use a Twilio webhook. The prospective or existing customer's number is logged in an Airtable Base which we'll be using to track all our enquries. Next, the message is sent to our AI Agent who can reply to the user and decide if an appointment booking can be made. The reply is made via SMS using Twilio. A scheduled trigger which runs every day, checks our chat logs for a list of prospective customers who have yet to book an appointment but still show interest. This list is sent to our AI Agent to formulate a personalised follow-up message to each lead and ask them if they want to continue with the booking. The follow-up interaction is logged so as to not to send too many messages to the customer. Requirements A Twilio account to receive customer messages. An Airtable account and Base to use as our datastore for enquiries. Cal.com account to use as our scheduling service. OpenAI account for our AI model. Customising this workflow Not using Airtable? Swap this out for your CRM of choice such as hubspot or your own service. Not using Cal.com? Swap this out for API-enabled services such as Acuity Scheduling or your own service.
By JimleukAutomate customer support issue resolution using AI text classifier
This n8n template is designed to assist and improve customer support team member capacity by automating the resolution of long-lived and forgotten JIRA issues. How it works Schedule Trigger runs daily to check for long-lived unresolved issues and imports them into the workflow. Each Issue is handled as a separate subworkflow by using an execute workflow node. This allows parallel processing. A report is generated from the issue using its comment history allowing the issue to be classified by AI - determining the state and progress of the issue. If determined to be resolved, sentiment analysis is performed to track customer satisfaction. If negative, a slack message is sent to escalate, otherwise the issue is closed automatically. If no response has been initiated, an AI agent will attempt to search and resolve the issue itself using similar resolved issues or from the notion database. If a solution is found, it is posted to the issue and closed. If the issue is blocked and waiting for responses, then a reminder message is added. How to use This template searches for JIRA issues which are older than 7 days which are not in the "Done" status. Ensure there are some issues that meet this criteria otherwise adjust the search query to suit. Works best if you frequently have long-lived issues that need resolving. Ensure the notion tool is configured as to not read documents you didn't intend it to ie. private and/or internal documentation. Requirements JIRA for issues management OpenAI for LLM Slack for notifications Customising this workflow Why not try classifying issues as they are created? One use-case may be for quality control such as ensuring reporting criteria is adhered to, summarising and rephrasing issue for easier reading or adjusting priority.
By JimleukAutomate competitor research with Exa.ai, Notion and AI agents
This n8n workflow demonstrates a simple multi-agent setup to perform the task of competitor research. It showcases how using the HTTP request tool could reduce the number of nodes needed to achieve a workflow like this. How it works For this template, a source company is defined by the user which is sent to Exa.ai to find competitors. Each competitor is then funnelled through 3 AI agents that will go out onto the internet and retrieve specific datapoints about the competitor; company overview, product offering and customer reviews. Once the agents are finished, the results are compiled into a report which is then inserted in a notion database. Check out an example output here: https://jimleuk.notion.site/2d1c3c726e8e42f3aecec6338fd24333?v=de020fa196f34cdeb676daaeae44e110&pvs=4 Requirements An OpenAI account for the LLM. Exa.ai account for access to their AI search engine. SerpAPI account for Google search. Firecrawl.dev account for webscraping. Notion.com account for database to save final reports. Customising the workflow Add additional agents to gather more datapoints such as SEO keywords and metrics. Not using notion? Feel free to swap this out for your own database.
By JimleukAutomate blog creation in brand voice with AI
This n8n template demonstrates a simple approach to using AI to automate the generation of blog content which aligns to your organisation's brand voice and style by using examples of previously published articles. In a way, it's quick and dirty "training" which can get your automated content generation strategy up and running for very little effort and cost whilst you evaluate our AI content pipeline. How it works In this demonstration, the n8n.io blog is used as the source of existing published content and 5 of the latest articles are imported via the HTTP node. The HTML node is extract the article bodies which are then converted to markdown for our LLMs. We use LLM nodes to (1) understand the article structure and writing style and (2) identify the brand voice characteristics used in the posts. These are then used as guidelines in our final LLM node when generating new articles. Finally, a draft is saved to Wordpress for human editors to review or use as starting point for their own articles. How to use Update Step 1 to fetch data from your desired blog or change to fetch existing content in a different way. Update Step 5 to provide your new article instruction. For optimal output, theme topics relevant to your brand. Requirements A source of text-heavy content is required to accurately breakdown the brand voice and article style. Don't have your own? Maybe try your competitors? OpenAI for LLM - though I recommend exploring other models which may give subjectively better results. Wordpress for blog but feel free to use other preferred publishing platforms. Customising this workflow Ideally, you'd want to "train" your agent on material which is similar to your output ie. your social media post may not get the best results from your blog content due to differing formats. Typically, this brand voice extraction exercise should run once and then be cached somewhere for reuse later. This would save on generation time and overall cost of the workflow.
By JimleukGenerate logos and images with consistent visual styles using Imagen 3.0
This n8n template allows you to use AI to generate logos or images which mimic visual styles of other logos or images. The model used to generate the images is Google's Imagen 3.0. With this template, users will be able to automate design and marketing tasks such as creating variants of existing designs, remixing existing assets to validate different styles and explore a range of designs which would have been otherwise too expensive and time-consming previously. How it works A form trigger is used to capture the source image to reference styles from and a prompt for the target image to generate. The source image is passed to Gemini 2.0 to be analysed and its visual style and tone extracted as a detailed description. This visual style description is then combined with the user's initial target image prompt. This final prompt is given to Imagen 3.0 to generate the images. A quick webpage is put together with the generated images to present back to the user. If the user provided an email address, a copy of this HTML page will be sent. How to use Ensure the workflow is live to share the form publicly. The source image must be accessible to your n8n instance - either a public image of the internet or within your network. For best results, select a source image which has strong visual identity as these will allow the LLM to better describe it. For your prompt, refer to the imagen prompt guide found here: https://ai.google.dev/gemini-api/docs/image-generationimagen-prompt-guide Requirements Gemini for LLM and Imagen model. Cloudinary for image CDN. Gmail for email sending. Customising this workflow Feel free to swap any of these out for tools and services you prefer. Want to fully automate? Switch the form trigger for a webhook trigger!
By Jimleuk