Browse 12,955 Workflows

By Amplence

Quick overview Turn vendor invoices from Gmail into approved QuickBooks bills. Claude reads each PDF, code checks it against your vendors, POs, log and QuickBooks, Slack approvers decide, and every outcome is logged, labelled and followed up. How it works A Gmail trigger catches each email with a PDF and calls this workflow once per email, so every invoice runs and waits for approval on its own. The PDF is saved to a Google Drive folder, then one Anthropic API call reads it and returns vendor, invoice number, dates, amounts, PO number, bank details and line items as fixed JSON. It reads your Vendors, Purchase orders and Invoice log tabs and searches QuickBooks Online for a bill with the same number. Code checks the vendor, sender domain, bank details, arithmetic, open PO balance and duplicates, then picks auto-approve, approve or hold. Small, clean, PO-matched invoices approve themselves. Others go to a Slack approver by direct message with a time limit, plus a second approver above your limit. Whoever raised the PO never approves its invoice. Every outcome is logged and posted to your AP channel. Approved invoices become QuickBooks bills with one line per invoice line. Changed bank details alert finance and never reply to the sender. Fixable problems get a reply asking the vendor for a corrected invoice, and Gmail labels show the result. Each weekday morning a digest lists approved bills coming due and invoices still waiting on a person. An error trigger posts any failure to Slack. Setup Create a Google Sheet with the tabs Vendors, Purchase orders and Invoice log, using the column headers in the canvas note, and add each vendor's QuickBooks vendor ID and expense account ID. Fill in the Settings node: sheet ID, Drive folder ID, Slack user and channel IDs, Gmail label IDs, QuickBooks company ID, and your approval limits. Connect Gmail, Google Drive, Google Sheets, Slack and QuickBooks Online credentials, and add a Header Auth credential named x-api-key with your Anthropic API key. Choose this workflow as its own error workflow in Workflow settings. Test from Each invoice starts here with the pinned sample (run it twice to see the duplicate hold), then unpin Read the invoice with Claude and activate. Requirements Gmail, Google Drive, Google Sheets, Slack and QuickBooks Online accounts, plus an Anthropic API key Customization Change limits and tolerances in Settings, add your own checks in Check and route, or point Create the bill in QuickBooks at Xero or another accounting API. Additional info Built by Amplence (amplence.com), an n8n automation agency. The checks and the design choices behind them: amplence.com/blog/n8n-invoice-approval-workflow. An n8n automation workflow template by Amplence.

By Weio

Quick overview This workflow runs weekly (or manually) to read a list of client websites from Google Sheets, check each site’s HTTPS and certificate status via weio.ai, write results back to the sheet, and send Slack and Gmail alerts for new or ongoing problems. How it works Runs every Monday at 08:00 (or when executed manually) and loads alert thresholds and the digest email address. Reads website rows from Google Sheets and keeps only rows with a non-empty website and not marked inactive. Calls the weio.ai HTTPS check API for each website (batched one at a time with a delay) to evaluate reachability, HTTPS redirect behavior, and certificate validity/expiry. Classifies each result as ok, warning, or critical, compares it to the previous status stored in the sheet, and timestamps the check. Updates the matching rows in Google Sheets with the latest severity and HTTPS status details. Sends an urgent Slack message for any critical issue that is new or has changed since the last run, and emails a single Gmail digest listing all non-ok sites. If the workflow execution fails, posts an error notification to Slack via an Error Trigger. Setup Create a Google Sheet with columns like website, active (set to "no" to skip), owner (optional), and output columns severity, https_status, https_detail, cert_expires, and checked_at. Add Google Sheets OAuth credentials and select the target document and sheet in both the read and update steps. Add a Slack credential and choose the channel for both the urgent alert and the workflow-failed notification. Add a Gmail OAuth credential for sending the email digest. Update the Settings values (email_to, warn_days, critical_days) to match your recipients and certificate warning windows, and (optionally) add an Authorization bearer token to the weio.ai request if you need higher throughput. Requirements Google Sheets, Slack and Gmail accounts connected as n8n credentials. No account is needed for the weio.ai HTTPS check: the free tier allows 5 checks a minute, and the workflow spaces requests 15 seconds apart (about 240 sites an hour). Customization Change warn_days and critical_days in Settings to move the expiry thresholds. For bigger lists, add a Header Auth credential (Authorization: Bearer ) on Check HTTPS and shorten the batch interval. Additional info API docs and keys: https://weio.ai/services/site-check-api.html?utm_source=n8n&utm_medium=templates&utm_campaign=https-watch . Built and maintained by Weio, Inc.; AI operators did most of the work. An n8n automation workflow template by Weio.

By JobsPipe - Dvir

Quick overview This workflow opens an n8n chat where an OpenAI-powered agent uses the JobsPipe MCP server to answer questions about who is hiring, returning job-posting-backed results with source links and optional monitoring signals. How it works Receives a message in an n8n chat trigger. Applies default search context such as your persona, default country, and the maximum number of jobs to return per search. Uses an OpenAI chat model and the last 10 chat messages as memory to interpret the user’s question. Calls JobsPipe MCP tools (for example, searching postings, fetching full job details, detecting company tech stacks, or creating monitoring signals) to gather live job data. Responds in the chat with results that cite the original job posting URLs and notes when no matches are found. Setup Create a free JobsPipe account at jobspipe.dev. In the JobsPipe MCP node, create an MCP OAuth2 credential named "JobsPipe MCP (OAuth)": Server URL https://mcp.jobspipe.dev, Resource https://mcp.jobspipe.dev/mcp, dynamic client registration on, then click Connect and sign in. An API key also works: switch to Header Auth with Authorization: Bearer jp_live_... Add an OpenAI credential named "OpenAI account" (or pick another chat model), then open each node once after importing so n8n attaches the credentials. Update the default persona, country, and max_jobs_per_search values to match your use case, then open the chat and start asking hiring-market questions. Requirements A JobsPipe account (the free tier includes 1,000 credits; each job returned costs 1 credit) and an OpenAI API key Customization Edit the agent's system prompt or the Set node to focus on a role, region or industry, raise max_jobs_per_search, or swap the chat model Additional info Ask the agent to watch a search and it can create a JobsPipe Signal that alerts you when new matching jobs appear. Docs: https://docs.jobspipe.dev/integrations/n8n. An n8n automation workflow template by JobsPipe - Dvir.

By Friedrich Bremer

Quick overview This workflow lets you send receipt photos to a Telegram bot, extracts key fields with OpenAI Vision, stores the image in Google Drive, and logs a German tax-ready expense row to Google Sheets with deductible amount and VAT handling, including optional voice-note refinements and duplicate detection. How it works Triggers on incoming Telegram messages or button callbacks and allows only chat IDs listed in the workflow’s Config. When you send a receipt photo, it downloads the image from Telegram and uses OpenAI (GPT-4o-mini vision) to extract structured receipt data such as merchant, date, gross/net, VAT, currency, and payment method. If the receipt is unreadable or missing required fields, it sends an error back to Telegram asking for a clearer photo. For readable receipts, it creates a descriptive filename, uploads the image to Google Drive, and calls OpenAI again to classify the expense into a cost center with German tax notes, deductibility, and VAT recoverability flags. It calculates deductible amounts and VAT fields in-code using German rules (for example, net vs. gross deduction base and VAT blocking cases), generates a tracking ID, and checks Google Sheets for likely duplicates. If a duplicate is suspected, it posts a Telegram warning with “Add anyway” or “Ignore”; otherwise it appends the row to the Ausgaben tab in Google Sheets and sends a confirmation message with buttons. When you tap buttons or reply with a voice note, it can prompt for context, transcribe the voice note with OpenAI Whisper, reclassify/update the existing Google Sheets row, or delete the row and edit the Telegram message accordingly. Setup Create a Telegram bot with BotFather, add the Telegram credentials in n8n, and start a chat with the bot so it can receive your messages. Add an OpenAI API credential (used for both GPT-4o-mini vision/classification and Whisper transcription). Add Google Sheets OAuth credentials and select a spreadsheet that contains an Ausgaben sheet with the columns expected by the workflow’s mapping. Add Google Drive OAuth credentials and choose the target Drive and folder where receipt images should be uploaded. In the ⚙️ Config node, set owner_chat_ids to your Telegram chat ID(s) and set sheet_url to the Google Sheet link used by the “View Sheet” buttons. An n8n automation workflow template by Friedrich Bremer.

By Stathis

Quick overview This workflow triggers on NUACOM calls, fetches the AI-generated transcript, and uses Google Gemini to extract structured recruiting insights. It writes the Q&A back to the NUACOM call notes and updates a HubSpot contact property with the candidate’s salary expectation. How it works Triggers when a NUACOM call event is received for the configured extension. Checks the call duration and continues only if it is longer than 5 seconds. Waits 30 seconds, then retrieves NUACOM’s AI-generated call analysis including the transcription. Sends the transcript to Google Gemini with instructions to answer 11 recruiting questions and return a JSON-structured result. Reformats the structured answers into a simple JSON object and adds a formatted note to the NUACOM call with the extracted details. Searches HubSpot for a contact whose phone number matches the call recipient and, if the salary expectation field is empty, updates the contact with the extracted salary expectation. Setup Add your NUACOM API credentials, set the extension to monitor, and ensure NUACOM AI Call Analysis is enabled for your calls. Configure access to Google Gemini (Google PaLM/Gemini credentials via n8n’s AI integration) for the AI Agent. Add your HubSpot OAuth2 credentials and ensure the custom contact property ain8n_salary_expectation exists (and adjust the property name if yours differs). An n8n automation workflow template by Stathis.

By Felix M

Quick overview YouTube removed the 'Upload Date: Last Hour' filter. This workflow fills that gap, running hourly to search YouTube via the RecentReborn Search API, deduplicating results, and logging them to Google Sheets, with Telegram alerts on failure. How it works Runs every hour on a schedule (or manually for testing). Reads your comma-separated search terms and splits them into individual queries with shared settings like upload window and video type. Calls the RecentReborn Search API for each query to retrieve recent YouTube video results uploaded within the last hour. Validates that each API response includes a results list and sends a Telegram message when the request fails or the response format is unexpected. Deduplicates videos by URL using workflow static data so only previously unseen videos from this run continue. Appends new videos to Google Sheets or updates existing rows by matching on the url column. Setup Create a RecentReborn API key and add it as an HTTP Bearer Auth credential used by the RecentReborn search request. Connect your Google Sheets OAuth credential and set the target spreadsheet and sheet tab for the append/update action. Connect a Telegram bot credential, set your Telegram chat ID in the workflow settings, and use it to receive error alerts. Update the search terms (comma-separated) and optional filters (upload window, type, duration) in the configuration step, then run once manually before enabling the hourly schedule. Requirements A RecentReborn API account (app.recentreborn.com) with an API key Additional info Visit https://recentreborn.com/developers for more tutorials and templates. An n8n automation workflow template by Felix M.

By Cristian Dinca

Quick overview This workflow replies to incoming Telegram messages using a Gemini agent that references your business settings, checks and books appointments in Google Calendar, logs each interaction to Google Sheets, and alerts the owner in Telegram when a human handoff is needed. How it works Triggers when a new message is received by your Telegram bot. Loads your business settings (name, language, knowledge base, appointment length, and owner chat ID) and passes the customer message to a Gemini agent with 12-turn conversation memory. When the customer asks to book, the agent checks real availability in Google Calendar and creates an event after the customer confirms a time. The agent appends a CRM-style log entry to Google Sheets with the time, customer, message, outcome, and whether a human is needed. The workflow removes the internal handoff token from the reply, sends the customer response back in Telegram, and notifies the owner in Telegram if the agent flagged the conversation for human follow-up. Setup Create and connect credentials for Telegram, Google Gemini, Google Calendar, and Google Sheets. In “Your settings”, fill in your business name, default language, appointment length, knowledge text (hours/services/prices/policies), and your owner Telegram chat ID. Set the Google Sheet URL and sheet name for logging, and ensure the sheet has columns for Time, Customer, Message, Outcome, and Needs Human. Confirm the Google Calendar to use (for example, the primary calendar) and test by messaging your Telegram bot to verify replies, booking, logging, and owner alerts. Requirements A Telegram bot token from BotFather (free) A Google account with Google Calendar and Google Sheets A Google Gemini API key (free tier is enough to start) Customization Swap the Telegram trigger and send nodes for WhatsApp Cloud API - the agent and its tools stay the same Rewrite the knowledge box in Your settings to serve a dental clinic, a garage or a restaurant Change the appointment length, or the 12-turn window in Conversation memory for longer chats. An n8n automation workflow template by Cristian Dinca.

By Mashrur Rahman

Quick overview This workflow captures a new client onboarding form submission with a proposal PDF, extracts key text, creates a Google Drive folder and a ClickUp folder/list, uses OpenAI to generate onboarding tasks, creates those tasks in ClickUp, then creates a Slack channel, sends a Slack welcome message and Gmail welcome email. How it works Receives a new submission from an n8n Form that includes client details and a proposal/scoping PDF. Extracts text from the uploaded PDF and trims it to a fixed length for downstream processing. Creates a new client folder in Google Drive and builds a shareable Drive folder URL. Creates a new folder and onboarding list in ClickUp for the client. Sends the extracted proposal text to OpenAI and parses the structured response into 10–30 task items with titles, descriptions, and due dates. Loops through the generated tasks and creates each one as a task in the ClickUp onboarding list. Creates a new public Slack channel for the client, posts a welcome message with the Drive link, and sends a welcome email via Gmail to the client. Setup Configure the n8n Form trigger basic auth and publish the form URL for submissions. Add Google Drive credentials and set the parent folder ID where new client folders should be created. Add ClickUp access token credentials and set the target Workspace/Space values so the workflow can create folders, lists, and tasks. Add an OpenAI credential and ensure the selected chat model (gpt-5-mini) is available in your account. Add Slack OAuth2 credentials with permission to create channels and post messages in your workspace. Add Gmail credentials for sending email and confirm the sender account is allowed to email your clients. An n8n automation workflow template by Mashrur Rahman.

By Shaharuk Anik

Quick overview This workflow captures Shopify checkout and order webhooks, logs activity to Google Sheets, and uses an Ollama-hosted Llama 3.2 chat endpoint to generate and send cart-recovery and post-purchase emails via Gmail, then emails a weekly recovery summary. How it works Receives Shopify webhook events for checkout creation/updates and order creation. For checkout events, logs the checkout details (email, product, price, recovery link, and cart token) to Google Sheets and starts the follow-up sequence only for new checkouts. After 1 hour, looks up the checkout by cart token in Google Sheets, skips if its status is already “Recovered,” and uses Llama 3.2 (via an HTTP call to an Ollama chat endpoint) to draft a reminder email that Gmail sends to the customer. After 24 more hours, re-checks the same Google Sheets row, skips if “Recovered,” generates a discount follow-up mentioning code CART10 with Llama 3.2, sends it via Gmail, and updates the row status to “Discount Sent.” After another 24 hours, re-checks the Google Sheets row again, skips if “Recovered,” generates a final urgency email with Llama 3.2, sends it via Gmail, and updates the row status to “Urgency Sent.” For order creation events, updates the matching Google Sheets row to “Recovered,” picks a random upsell product from a built-in catalog, generates a thank-you email with Llama 3.2, and sends it via Gmail. Every Monday at 9am, reads all tracker rows from Google Sheets, calculates weekly recovered revenue and recovery rate, and emails the summary via Gmail. Setup Create Shopify webhooks for checkout creation/updates and order creation and point them to this workflow’s webhook URL. Create a Google Sheet with columns for Checkout Token, Email, Product, Price, Status, Recovery Link, Created At, and Last Updated, then add a Google Sheets OAuth2 credential and replace YOUR_GOOGLE_SHEET_ID_HERE in all Google Sheets steps. Add a Gmail OAuth2 credential, set the weekly summary recipient email address, and ensure your Gmail account is allowed to send to your customers. Provide an Ollama (or compatible) chat endpoint that serves the llama3.2 model and update the HTTP request URL if it is not reachable at http://host.docker.internal:11434/api/chat. If you use the discount follow-up, create the CART10 discount code in Shopify (or update the prompts/code to match your discount code and branding). Requirements Shopify store with admin access Google account with Sheets and Gmail access Ollama running locally with Llama 3.2 (or any OpenAI-compatible chat API) Customization Swap the local Llama model for GPT-4o or Claude by updating the HTTP Request nodes' endpoint and body Edit the product list in the Pick Upsell node and the store name in the Code nodes Adjust the three Wait nodes, or change the discount code and email tone in the HTTP Request prompts. An n8n automation workflow template by Shaharuk Anik.

By Tyler Eastman

Quick overview This workflow runs every weekday morning, searches You.com for the latest AI news by topic, scores each article with TypeSafe’s Jev model, and posts a short cited brief per topic to a Slack channel. How it works Runs every weekday at 8am based on your n8n instance timezone. Splits the configured list of topics into individual searches and uses You.com Search to fetch up to 15 articles from the past day for each topic. Normalizes each result into a consistent article record and skips URLs that were already processed in previous executions. Sends each new article to TypeSafe’s Jev API to judge whether it reports a real development, is mainly about the topic, and how material it is for the configured team. Filters to keep only articles that meet the development, topicality, and materiality thresholds, then groups the remaining items by topic. Uses You.com Research to write a 1–4 bullet morning brief per topic using the grouped article list, formats citations as Slack links, and posts the brief to the selected Slack channel. Setup Install the You.com node (@youdotcom-oss/n8n-nodes-youdotcom), available on n8n Cloud and self-hosted. Add You.com API credentials and select them in both You.com nodes (Search and Research). Create an HTTP Header Auth credential for the Jev request with Authorization: Bearer . Add Slack credentials and set the target channel (the template posts to the channel named ai-news). Update the Config values (topics, team description, caresAbout list, and materialityThreshold) to match what your team wants to track. Additional info For more information and evals, see the repo in Github: https://github.com/youdotcom-oss/n8n-ai-brief-workflow. An n8n automation workflow template by Tyler Eastman.

By Sam

Quick overview Paste an article's address; this workflow lists its dated figures and, for each, a later-dated figure from the web, quoted word for word from its source, side by side for you to review. Self-hosted n8n only. How it works A form takes the article's address and, optionally, a region. The DuckDuckGo node reads the article. The model lists up to five figures that state the year they refer to; code keeps only those whose sentence is really in the article with the value and the year, and drops forecasts. One DuckDuckGo search per figure reads the top five pages, with a short wait between searches. The model proposes later figures; code keeps a candidate only if its quote is on the source page, holds the value and a later year close together, and is not a forecast. A different unit is flagged. The form shows a review table: the article's sentence next to the later one, with its source. Every way the run can end shows a page. Setup Self-hosted n8n: install n8n-nodes-duckduckgo-search under Settings → Community nodes. Searching and reading pages need no API key or account. Add a credential to the chat-model node, OpenRouter model (Gemini 3.8 Flash by default). Any other chat-model node can replace it; choose a model that returns JSON reliably. Activate the workflow, open the form's production URL and paste an article's address. The region is optional and keeps the searches to one country or market. Requirements Self-hosted n8n only: community nodes cannot be installed on n8n Cloud The community node n8n-nodes-duckduckgo-search An API key for the AI model: OpenRouter by default, or any provider with an n8n chat-model node No API key or account is needed for DuckDuckGo Customization Lower the limit of five figures (in Check the figures and the first prompt) for shorter runs; each figure is one DuckDuckGo search Set a region in the form to keep searches to one country or market Swap the model node for any other chat-model node Keep the wait between searches: DuckDuckGo limits requests per IP address Send the table by email or Slack instead of showing it on the form page Additional info Suggested title: "Find outdated statistics in an article with DuckDuckGo". The AI model is a swappable part (any chat-model node works), so the title names only the search node. The canvas preview does not render community nodes, so the full canvas is in the screenshot at the top of the description. The table offers candidates for review, not corrections: a quote shown is word for word on its source page, which does not prove the figure is right or measures the same thing. Source and notes: https://github.com/samnodehi/n8n-nodes-duckduckgo/tree/main/docs/examples. An n8n automation workflow template by Sam.

By pdfRest API Toolkit

Quick overview Upload a PDF containing multiple documents, use AI to suggest document groups, and review titles, categories, and page ranges in a guided form. pdfRest creates separate PDFs and a downloadable ZIP with a page manifest, with optional Google Drive filing. How it works Receives a PDF through an n8n upload form and verifies that exactly one valid PDF was provided. Uses pdfRest to inspect the PDF, run OCR to create a searchable master copy, and split the master into individual page PDFs. Uses pdfRest to convert each page to Markdown, then compiles a full-page inventory while preserving original page numbers and flagging low-text pages. Sends the combined extracted text to OpenAI (via the n8n LangChain Information Extractor) to propose contiguous document boundaries with document type, title, counterparty, and evidence-based reasoning. Presents editable titles, categories, and start/end page numbers for each proposed document. Reviewers can exclude groups or add up to two additional groups, then approve or cancel. Checks that every page belongs to exactly one included group. Invalid ranges receive clear correction messages, with edits preserved across up to three review submissions. Validated groups are sorted by start page, split into separate PDFs, and given descriptive filenames. Optionally uploads the split PDFs to Google Drive, then builds a page-to-document manifest, compresses the PDFs and manifest into a ZIP, and returns it as a download. Setup Add a pdfRest API credential and connect it to all pdfRest nodes (PDF info, OCR, split pages, convert to Markdown, and split approved documents). Add an OpenAI API credential in the AI model node and select a model available to your account. Review and adjust the Settings values (maxPages, maxTextCharacters, and ocrLanguages) to match your expected batch sizes and languages. If you enable Google Drive filing (saveToDrive = true), add a Google Drive credential and set outputFolderId to a real destination folder ID. Requirements Requires the pdfRest API Toolkit node, pdfRest and OpenAI or Anthropic credentials, and Google Drive credentials only when optional Drive filing is enabled. An n8n automation workflow template by pdfRest API Toolkit.