Best n8n Templates & Workflows to Start With in 2026
If you are new to n8n, the blank canvas is the hardest part. The fastest way past it is to start from a template - a ready-made node graph you import, connect to your own credentials, and adapt. Below are nine representative workflows drawn from n8n's own examples, covering the patterns most people reach for first: AI agents, web scraping, chatbots, data conversion, enrichment, backups, APIs, and retrieval. These are starting points, not a ranked benchmark.
The Starter Set at a Glance
The table below lists all nine templates. Click any column header to sort, and click a template name to jump to its detail. The number is a suggested learning order, not a measured score - every one of these is a legitimate place to begin.
| #▲ | Template▲▼ | Use Case▲▼ | Key Nodes / Concepts▲▼ | Best For▲▼ |
|---|---|---|---|---|
| 1 | AI Agent Chat | AI | AI Agent node, chat model, tools | AI beginners |
| 2 | Scrape & Summarize | AI | HTTP Request, Summarization Chain | Research / monitoring |
| 3 | Telegram AI Chatbot | AI | Telegram trigger, AI Agent, memory | Support / personal bots |
| 4 | JSON to Excel | Data | Convert to File, Edit Fields | Reporting / ops |
| 5 | Company Enrichment | AI / Data | HTTP Request, Information Extractor | Sales / marketing |
| 6 | Back Up to GitHub | DevOps | Schedule trigger, n8n API, GitHub | Teams / admins |
| 7 | Create an API Endpoint | Integration | Webhook, Respond to Webhook | Developers |
| 8 | Join Datasets | Data | Merge node, matching keys | Analysts |
| 9 | RAG with Pinecone | AI | Embeddings, Pinecone vector store | AI / knowledge bases |
This is a curated starting set, not a benchmark leaderboard. No vendor paid for placement, and Tech Jacks Solutions has no affiliate relationship with n8n.
#1 AI Agent Chat
This is the canonical "hello world" of n8n's AI features, and the one we recommend opening first. A chat trigger feeds messages into the AI Agent node, which is connected to a chat model (you can pick OpenAI, Anthropic, Google Gemini, Mistral, Ollama, and others as sub-nodes) and optionally to tools. Unlike a plain text-generation call, the AI Agent is goal-oriented: it reasons about what to do, can call attached tools, and runs multi-step before replying.
The template teaches the core building block of every n8n AI workflow: the root agent node plus its sub-nodes (the chat model, memory, and tools all attach beneath it). Once you understand this cluster pattern, the more advanced templates further down this list stop looking intimidating. n8n's AI features are built on LangChain, and the agent supports several types, including Conversational, ReAct, Plan and Execute, and Tools.
What it uses: AI Agent node (root) + a chat model sub-node + optional Calculator, Wikipedia, SerpApi, or MCP Client tools.
Who it is for: Anyone new to AI in n8n. Start here, swap in your own model credentials, and you have a working assistant in minutes.
#2 Scrape and Summarize Webpages with AI
A practical first AI workflow that does real work. The HTTP Request node fetches a webpage, the content is passed through extraction, and a Summarization Chain condenses it into a short brief. Point it at a competitor's blog, a news source, or a documentation page, and you get a digest without reading the whole thing.
This template is a good lesson in combining a plain integration node (HTTP Request, which can reach essentially any URL) with an AI chain. It also shows why n8n's per-execution model is friendly to this pattern: scraping and summarizing one page is a single execution regardless of how many steps the workflow has internally.
What it uses: HTTP Request node to fetch the page, plus a Summarization Chain (one of n8n's built-in AI chains) connected to your chat model of choice.
Who it is for: Researchers, analysts, and anyone who monitors web sources and wants the gist instead of the full text. Pair it with a schedule trigger to run it automatically.
#3 Telegram AI Chatbot
This template puts an AI agent behind a real messaging surface. A Telegram trigger fires whenever someone messages your bot, the message goes to an AI Agent node, and the reply is sent back through Telegram. Add a memory sub-node (Simple Memory, Postgres, Redis, and others are supported) and the bot remembers the conversation across turns.
It is the natural next step after the AI Agent Chat template, because it shows how to attach an agent to an external trigger and channel rather than n8n's built-in chat window. The same shape works for Slack, Discord, or a website widget - you swap the trigger and the response node, and the agent in the middle stays the same.
What it uses: Telegram trigger and Telegram send nodes, an AI Agent node with a chat model, and a memory sub-node to maintain context.
Who it is for: Anyone building a support bot, a personal assistant, or an internal helper that lives where people already chat. A good intro to conversational memory.
#4 Convert JSON to Excel
Not every useful template involves AI, and this one is a reminder that n8n is a general automation tool first. It takes structured JSON - the format almost every API returns - and produces a spreadsheet file. The Convert to File node handles the transformation to XLSX (or CSV), and you can shape the data first with the Edit Fields node so the columns come out clean.
This is the workflow people quietly run the most. An API gives you JSON; a stakeholder wants a spreadsheet. The template closes that gap in two or three nodes and is a clear demonstration of how n8n moves data between formats without writing parsing code.
What it uses: An input source returning JSON, the Edit Fields node to select and rename columns, and the Convert to File node to emit Excel or CSV.
Who it is for: Operations, finance, and anyone who regularly turns API output or database rows into a report a colleague can open.
#5 Company Enrichment from a Website
Give this workflow a company's website and it returns structured details about that company. The HTTP Request node pulls the site content, and the Information Extractor - one of n8n's AI components - reads the unstructured page text and returns clean fields such as industry, location, or description. The output is structured data you can write to a CRM or spreadsheet.
It is a strong example of the "unstructured in, structured out" pattern that makes AI genuinely useful in automation. Where the Summarization Chain gives you prose, the Information Extractor gives you a predictable schema, which is exactly what downstream nodes need to act on.
What it uses: HTTP Request to fetch the site, the Information Extractor AI node to pull defined fields, and an output node (Sheets, a CRM, or a database) to store the result.
Who it is for: Sales and marketing teams building lead lists, and anyone who needs to turn a pile of URLs into a structured table.
#6 Back Up Workflows to GitHub
This is housekeeping that pays for itself the first time someone breaks a workflow. On a schedule trigger, the workflow calls n8n's own API to export every workflow definition as JSON, then commits those files to a GitHub repository. You get version history, diffs, and a recovery point for your automations - the same discipline you would apply to code.
It is a useful template to study because it shows n8n automating itself. It also makes the case for treating your workflows as artifacts worth protecting, especially on self-hosted instances where there is no managed backup behind you.
What it uses: A schedule trigger, the n8n node (or HTTP Request against the n8n API) to list and export workflows, and the GitHub node to commit the files.
Who it is for: Teams and administrators running more than a handful of workflows, and anyone self-hosting who wants a real disaster-recovery story.
#7 Create an API Endpoint
n8n is not only a consumer of APIs - it can serve one. The Webhook trigger gives your workflow a public URL that accepts incoming requests, you process the request with whatever nodes you need in the middle, and the Respond to Webhook node returns a custom response to the caller. The result is a working HTTP endpoint backed by a visual workflow.
This template reframes how people think about n8n. Instead of "a tool that runs on a schedule," it becomes "a place I can stand up a small service." You can build a lightweight microservice, a form handler, or an integration shim without provisioning a server or writing boilerplate routing code.
What it uses: A Webhook trigger to receive the request, any processing nodes (including the Code Node for custom JavaScript or Python), and the Respond to Webhook node to send the reply.
Who it is for: Developers who need a quick endpoint, and teams gluing systems together that expect to call a URL rather than poll on a schedule.
#8 Join Datasets
Real workflows rarely pull from a single source. This template takes two datasets - say, customers from a database and orders from an API - and combines them on a matching key using the Merge node. It is the n8n equivalent of a SQL join or a spreadsheet VLOOKUP, done visually and repeatably.
Learning the Merge node early removes a common frustration: people often try to chain two sources together linearly and lose data, when what they actually need is a join. This template demonstrates matching by field, combining by position, and appending - the three things the Merge node does that solve most multi-source problems.
What it uses: Two input branches (each fetching a dataset), and the Merge node configured to combine them by a matching key.
Who it is for: Analysts and anyone who reconciles data across systems. It is the bridge from single-source automations to genuinely useful ones.
#9 RAG with Pinecone
This is the most advanced pattern on the list and the one that turns an AI agent into something that knows your content. Retrieval-augmented generation (RAG) stores your documents as embeddings in a vector database, then retrieves the most relevant chunks at query time and feeds them to the model. n8n supports several vector stores - this pattern uses Pinecone, with Qdrant, Chroma, Supabase, PGVector, Weaviate, and others available as alternatives.
There are really two workflows here: an ingestion run that chunks and embeds your documents into Pinecone, and a query workflow where an AI agent uses the Pinecone vector store as a retrieval tool. We put it last deliberately - it builds on the agent, tool, and chain concepts from the earlier templates, so it makes the most sense once those are familiar.
What it uses: An embeddings sub-node, the Pinecone vector store node for both insert and retrieval, and an AI Agent node that queries the store as a tool.
Who it is for: Teams building a chatbot over internal documentation, a knowledge base assistant, or any AI that must answer from a specific, private corpus rather than general training data.
How We Curated This List
n8n's library is large - the company reports somewhere between 800 and 900-plus templates, a figure that is vendor-reported and spans a range, so treat it as a sense of scale rather than a precise count. With that many options, "best" cannot mean a single ranked order. Instead, we chose nine workflows that meet three tests:
- Grounded in n8n's own examples: Each template here is a pattern n8n itself highlights in its homepage and documentation, not something we invented.
- Teaches a transferable concept: Every entry introduces a building block - the AI Agent node, the Merge node, webhooks, vector stores - that you will reuse across dozens of future workflows.
- Useful on day one: Each one does something real once you connect your own credentials, rather than being a toy demo.
What this list is not: A benchmark or a ranking by popularity, speed, or quality. n8n does not publish a measured ranking of its templates, so we do not pretend one exists. The numbers are a suggested learning order. Browse the full library inside the n8n editor or on the n8n Workflows page to find templates specific to your tools.
No vendor paid for inclusion. Tech Jacks Solutions has no affiliate relationship with n8n. Template examples and node names were verified against n8n's official documentation as of June 2026.
Frequently Asked Questions
How many templates does n8n have?
n8n reports a library of 800-plus to 900-plus workflow templates. That figure is vendor-reported and spans a range, so treat it as a rough indication of scale rather than an audited count. Templates are browsable inside the n8n editor or on the Workflows page at n8n.io.
Are these the best n8n templates by some benchmark?
No. This is a curated set of representative starting points grounded in the example workflows n8n itself cites. They are not ranked by any measured benchmark, and no vendor paid for placement. The numbering reflects a suggested learning order, not a score.
Where do I find n8n templates?
Templates live in two places: inside the n8n editor, where you can browse and import them directly into your instance, and on the n8n.io Workflows page. Importing a template drops a ready-made node graph onto your canvas that you then connect to your own credentials.
Do I need to know how to code to use n8n templates?
No. Most templates work with no code - you connect credentials and adjust parameters in a visual editor. When you do need custom logic, the Code Node runs JavaScript or Python, and the HTTP Request node reaches any REST or GraphQL API. That code-plus-visual mix is what n8n means by "the flexibility of code with the speed of no-code."
What is an AI Agent in n8n?
The AI Agent node is built on LangChain and handles goal-oriented, multi-step tasks rather than a single text completion. It selects and executes actions using attached tools and a chat model. n8n offers several agent types, including Conversational, ReAct, Plan and Execute, and Tools.
Video Resources
Go Deeper
Resources from across Tech Jacks Solutions