OpenAI workflows
630 results — all source-linked n8n references
Audit multi-page CRO with GPT-4o-mini, Google Sheets, and Gmail alerts
Description Submit up to 3 landing page URLs using a simple form and the workflow audits each one automatically. For every page, GPT-4o-mini scrapes the content, scores it across five dimensions, identifies the top conversion issue, and generates five specific page fixes. Every result is logged to Google Sheets and pages that score below your alert threshold trigger an automatic Gmail alert with the full breakdown. Built for marketing managers, agency owners, and growth teams who need structured, repeatable conversion audits without spending hours reviewing pages manually. What This Workflow Does Scrapes and cleans each page automatically** — Fetches the raw HTML for every URL submitted, strips all scripts and styles, and prepares clean text for AI analysis Scores pages across five CRO dimensions** — GPT-4o-mini rates each page on overall conversion potential, CTA strength, trust score, and message clarity on a 1–10 scale Generates a letter grade per page** — Assigns A+ through F based on the overall score so you can rank pages at a glance Produces five page-specific fixes** — Every fix references actual content from the page and is scoped to be doable within 1–2 days — no generic
Optimize blog and product pages for Google AI Overviews with GPT-4 and Sheets
This workflow automatically monitors blog posts or product pages for new/updated content, analyzes Google AI Overviews to identify optimization gaps, auto-generates optimized titles, summaries, and schema markup, and applies these improvements to enhance visibility in AI search results. Who's it for • SEO managers optimizing for AI-driven search results • Content teams publishing 5+ articles per week • E-commerce teams managing product page visibility • Digital marketers tracking AI Overview presence How it works / What it does Detects new or updated blog posts / product pages via webhook or schedule Fetches raw content and existing metadata Queries Google Search API to analyze AI Overview coverage gaps AI generates optimized title, meta summary, and JSON-LD schema markup Pushes improvements back to CMS or target endpoint Logs all changes and scores to Google Sheet tracker How to set up Import this workflow Set up credentials (Google Search API, OpenAI/Anthropic, CMS/HTTP, Google Sheets) Update your site URL and content preferences Activate workflow Requirements • Webhook endpoint or CMS trigger • Google Custom Search API key • OpenAI / Anthropic / Grok API • CMS REST API or target
SOrchestrate credit onboarding checks with GPT-4o, Airtable, Gmail and Slack
How It Works This workflow automates credit operations onboarding by running KYC verification, credit bureau checks, identity validation, and sanctions screening through a single AI-powered agent. Built for credit operations teams, compliance officers, and fintech platforms, it eliminates manual eligibility reviews that are slow and error-prone. Triggered via webhook, the Credit Operations Agent orchestrates all verification tools simultaneously, then routes customers by eligibility status, eligible, ineligible, pending documentation, or compliance escalation. Each path prepares structured data stored in Airtable, triggers appropriate follow-up actions (email, Slack alerts), and logs a full audit trail. A final formatted response is returned to the originating system, closing the loop end-to-end with no manual handoffs. Setup Steps Set webhook URL and connect Credit Operations webhook node to your intake system. Add OpenAI API key to the OpenAI Chat Model node. Configure KYC, Credit Bureau, Identity, and Sanctions tool credentials. Add Gmail OAuth2 and Slack bot token for notification nodes. Connect Airtable API key; set base/table IDs for eligible and ineligible customer stores. P
Monitor and optimize carbon emissions for ESG reporting with GPT-4o, Slack and Sheets
How It Works This workflow automates end-to-end carbon emissions monitoring, strategy optimisation, and ESG reporting using a multi-agent AI supervisor architecture in n8n. Designed for sustainability managers, ESG teams, and operations leads, it eliminates the manual effort of tracking emissions, evaluating reduction strategies, and producing compliance reports. Data enters via scheduled pulls and real-time webhooks, then merges into a unified feed processed by a Carbon Supervisor Agent. Sub-agents handle monitoring, optimisation, policy enforcement, and ESG reporting. Approved strategies are auto-executed or routed for human sign-off. Outputs are consolidated and pushed to Slack, Google Sheets, and email, keeping all stakeholders informed. The workflow closes the loop from raw sensor data to actionable ESG dashboards with minimal human intervention. Setup Steps Connect scheduled trigger and webhook nodes to your emissions data sources. Add credentials for Slack (bot token), Gmail (OAuth2), and Google Sheets (service account). Configure the Carbon Supervisor Agent with your preferred LLM (OpenAI or compatible). Set approval thresholds in the Check Approval Required node. Map Googl
Monitor and optimize carbon emissions for ESG with GPT‑4o, Slack and Google Sheets
How It Works This workflow automates end-to-end carbon emissions monitoring, strategy optimisation, and ESG reporting using a multi-agent AI supervisor architecture in n8n. Designed for sustainability managers, ESG teams, and operations leads, it eliminates the manual effort of tracking emissions, evaluating reduction strategies, and producing compliance reports. Data enters via scheduled pulls and real-time webhooks, then merges into a unified feed processed by a Carbon Supervisor Agent. Sub-agents handle monitoring, optimisation, policy enforcement, and ESG reporting. Approved strategies are auto-executed or routed for human sign-off. Outputs are consolidated and pushed to Slack, Google Sheets, and email, keeping all stakeholders informed. The workflow closes the loop from raw sensor data to actionable ESG dashboards with minimal human intervention. Setup Steps Connect scheduled trigger and webhook nodes to your emissions data sources. Add credentials for Slack (bot token), Gmail (OAuth2), and Google Sheets (service account). Configure the Carbon Supervisor Agent with your preferred LLM (OpenAI or compatible). Set approval thresholds in the Check Approval Required node. Map Googl
Monitor news sentiment trends with GPT-4o-mini, MrScraper, Slack and Notion
Description This n8n template automatically monitors news sources daily, analyzes article sentiment using AI, and delivers structured intelligence reports to your team — all without any manual reading. It uses MrScraper to discover and extract articles, GPT-4o-mini to score sentiment and flag urgent issues, and delivers results to both Notion (for archiving) and Slack (for real-time alerts and daily digests). Whether you're tracking brand reputation, monitoring a competitor, or staying on top of industry trends, this workflow turns the open web into a fully automated radar system that runs every morning before your team starts their day. How It Works Phase 1 – Trigger & Config:** A Schedule Trigger fires daily at 8AM. The workflow reads your list of target news source URLs from a Google Sheet, then loops through each source one by one. Phase 2 – URL Discovery:** For each news source, the Map Agent crawls the page and extracts individual article URLs. URLs are filtered using include patterns to keep only actual article links, deduplicated, and capped at your configured maxArticles limit. Phase 3 – Article Extraction:** Each article URL is processed by the General Agent, which extrac
COptimize n8n workflow JSON using Azure OpenAI GPT-4o-mini
📊 Description Automate the optimization of your n8n workflows using AI-powered analysis and restructuring 🤖. This workflow receives any existing n8n JSON via webhook, analyzes it using Azure OpenAI, and returns a fully optimized, import-ready workflow file. It improves execution speed, reduces redundant nodes, enhances API efficiency, and ensures structural cleanliness — all while preserving the original logic. Perfect for automation builders who want cleaner, faster, and production-ready workflows instantly. 🚀 What This Template Does 📥 Receives an n8n workflow JSON via a secure POST webhook. 🔍 Extracts and validates the workflow body to ensure proper structure. 🤖 Sends the workflow to Azure OpenAI for performance and structural optimization. 🧹 Strips markdown, parses, and validates the AI-returned JSON. 📄 Converts the optimized workflow into a downloadable .json file. 📤 Returns the optimized workflow file directly to the caller. Key Benefits ✅ Automatically optimizes workflow performance and execution speed ✅ Reduces redundant nodes and unnecessary API calls ✅ Ensures clean, structured, import-ready n8n JSON ✅ Preserves original business logic and integrations ✅ Eliminate
Enrich Google Maps business leads using Apify, OpenAI, Jina AI, Telegram and Sheets
Automate your B2B lead generation by scraping Google Maps, generating AI business summaries, and extracting hidden contact emails from websites, all triggered via Telegram. Why Use This Workflow? Time Savings: Reduces lead research time from 4 hours of manual searching to 5 minutes of automated execution per 50 leads. Cost Reduction: Replaces expensive monthly lead database subscriptions with a pay-as-you-go model using Apify and OpenAI. Error Prevention: Uses AI to deduplicate results and ensure company summaries are professional and consistent for your CRM. Scalability: Allows you to trigger massive scraping tasks from your mobile phone via Telegram while the backend handles the heavy lifting. Ideal For Sales Development Reps (SDRs):** Rapidly building targeted lists of local businesses for cold outreach or door-knocking campaigns. Marketing Agencies:** Identifying new businesses in specific sectors (e.g., "Dentists in Paris") to offer SEO or advertising services. Real Estate Investors:** Finding specific commercial properties or business types in a geographic area to identify investment opportunities. How It Works Trigger: The workflow starts when you send a message to your Tele
Generate M&A due diligence PDF reports with LlamaIndex, OpenAI, Pinecone, and S3
Streamline M&A due diligence with AI. This n8n workflow automatically parses financial documents using LlamaIndex, embeds data into Pinecone, and generates comprehensive, AI-driven reports with GPT-5-mini, saving hours of manual review and ensuring consistent, data-backed insights. Why Use This Workflow? Time Savings: Reduces manual document review and report generation from days to minutes. Cost Reduction: Minimizes reliance on expensive human analysts for initial data extraction and summary. Error Prevention: AI-driven analysis ensures consistent data extraction, reducing human error and oversight. Scalability: Effortlessly processes multiple documents and deals in parallel, scaling with your business needs. Ideal For Investment Analysts & Private Equity Firms:** Quickly evaluate target companies by automating the extraction of key financials, risks, and business models from deal documents. M&A Advisors:** Conduct preliminary due diligence efficiently, generating comprehensive overview reports for clients without extensive manual effort. Financial Professionals:** Accelerate research and analysis of company filings, investor presentations, and market reports for critical decision
Summarize YouTube videos in Slack using AssemblyAI transcription and OpenAI
🧠 How it works This workflow lets users generate structured summaries from YouTube videos directly inside Slack using n8n, AssemblyAI, and OpenAI. When a user submits a YouTube link via a Slack slash command, the workflow extracts the video ID and validates the video duration. Videos longer than the supported limit are stopped early with a clear message sent back to Slack. For valid videos, the workflow downloads the video audio as an MP3 file and sends it to AssemblyAI for transcription. Once the transcription is complete, the transcript is passed to an AI model to generate a structured summary. The final result includes a concise TL;DR, key takeaways, and notable quotes, which are formatted and posted back to Slack asynchronously using the original response URL. ⚙️ Features • Triggers from a Slack slash command with a YouTube link • Validates video length before processing (maximum 10 minutes) • Downloads YouTube audio as MP3 using RapidAPI • Transcribes audio using AssemblyAI • Generates structured summaries (TL;DR, key takeaways, notable quotes) • Posts the summarized result back to Slack asynchronously 💡 Use cases & expected outcomes • Educational YouTube videos → Receive a
WScore HubSpot deal conversion risk with OpenAI and Slack alerts
How it works This workflow runs daily to review all active deals and evaluate their likelihood of closing successfully. It enriches deal data with recent engagement activity and applies AI-based behavioral scoring to predict conversion probability. High-risk or stalled deals are flagged automatically. Actionable alerts are sent to the sales team, and all analysis is logged for forecasting and tracking. Step-by-step Trigger and fetch deals** Schedule Trigger – Runs the workflow automatically at a fixed time each day. Get Active Deals from HubSpot – Retrieves all open, non-closed deals with key properties. Formatting Data – Normalizes deal fields such as value, stage, age, contacts, and activity dates. Enrich deals with engagement data** If – Filters only active deals for further processing. Loop Over Items – Processes each deal individually. HTTP Request – Fetches engagement associations for the current deal. Get an engagement – Retrieves detailed engagement records from HubSpot. Extracts Data – Structures engagement content, timestamps, and metadata for analysis. Analyze risk, alert, and store results** OpenAI Chat Model – Provides the language model used for analysis. AI Agent – E
Discover buying-intent leads on Twitter and Instagram with GPT-4o-mini and send summaries to Slack and Notion
📊 Description Automate B2B lead discovery by identifying high-intent prospects directly from Reddit discussions using AI-powered intent analysis. 🎯🤖 This workflow scans Reddit for conversations related to CRM and marketing automation tools, analyzes snippets to detect buying intent, identifies relevant decision-makers on LinkedIn, enriches contact details via RocketReach, and logs qualified leads into Google Sheets. Running every three hours, it ensures your sales team never misses fresh outbound opportunities without manual research. 🚀📊 🔁 What This Template Does 1️⃣ Runs automatically every 3 hours to search Reddit for tool-related discussions. ⏰ 2️⃣ Extracts Reddit snippets, links, and highlighted keywords from search results. 🔍 3️⃣ Uses AI to classify buying intent as High, Medium, or Low. 🤖 4️⃣ Identifies the core problem and suggests a safe, non-salesy outreach angle. 💬 5️⃣ Filters only High and Medium intent opportunities. 🚦 6️⃣ Searches LinkedIn for matching decision-makers based on role and seniority. 👥 7️⃣ Enriches lead profiles with emails and company data using RocketReach. 📇 8️⃣ Saves qualified leads into Google Sheets with deduplication logic. 📊 9️⃣ Sends
Audit web pages for SEO issues with Decodo, GPT-4 and Google Sheets
What this workflow does This workflow automatically audits web pages for SEO issues and generates an executive-friendly SEO report using AI. It is designed for marketers, founders, and SEO teams who want fast, actionable insights without manually reviewing HTML, meta tags, or SERP data. The workflow: Reads URLs from Google Sheets Scrapes page content using Decodo (reliable scraping, even on protected sites) Extracts key SEO elements (title, meta description, canonical, H1/H2, visible text) Uses an AI Agent to analyze the page and generate: Overall SEO status Top issues Quick wins Title & meta description recommendations Saves results to Google Sheets Sends a formatted HTML executive report by email (Gmail) Who this workflow is for This template is ideal for: SEO consultants and agencies SaaS marketing teams Founders monitoring their landing pages Content teams doing SEO quality control It focuses on on-page SEO fundamentals, not backlink analysis or technical crawling. Setup (step by step) 1. Google Sheets Create an input sheet with one URL per row Create an output sheet to store SEO results 2. Decodo Add your Decodo API credentials The URL is automatically taken from the input she
Automate Gmail Organization with AI-Powered Email Classification and Smart Labeling
📧 AI-Powered Gmail Auto-Labeling with Smart Classification This n8n workflow automatically organizes your Gmail inbox by intelligently categorizing incoming emails and applying custom labels using AI-powered sentiment analysis. Say goodbye to manual email sorting and hello to an organized inbox! Use Cases Business Email Management**: Automatically sort partnership inquiries, meeting requests, and promotional emails Sales Team Automation**: Identify and label cold outreach and high-priority leads Executive Assistants**: Filter important meetings and urgent matters for busy professionals Marketing Teams**: Separate promotional content from genuine business communications Good to Know The workflow runs every minute to check for new emails Each email is processed only once - already labeled emails are automatically skipped Uses OpenAI's GPT model for accurate email classification (API costs apply) Processes emails in batches to handle multiple incoming messages efficiently How It Works Gmail Trigger continuously monitors your inbox for new emails every minute Switch Node checks if emails already have labels (Other, Promotion, or Meeting) to avoid duplicate processing Loop Node process
Automate Lead Response with GPT and Human Approval to Airtable CRM
AI-Assisted Lead Follow-Up With Human Approval This workflow automates your lead response process from end to end. When someone submits your n8n Form, the workflow generates an AI-written follow-up email, sends that draft to your sales team for approval, and then sends the approved email to the lead or marks it as needing revision. All lead details, drafts, approval decisions, and timestamps are stored in Airtable. Ideal For Teams that want AI to draft emails while keeping a human approval step Businesses receiving inbound inquiries that need fast, consistent responses Users building a simple form → email → CRM workflow Teams that want a record of all drafts and approval outcomes What This Template Provides AI-generated follow-up email drafts Human approval flow, using Approve and Reject links Automatic Airtable logging of leads, drafts, and statuses Fully automated pipeline triggered by a simple form submission Setup Steps (5–10 minutes) Connect Your Credentials Gmail Airtable OpenAI Create Your Airtable Table Use the following fields: Name Email Phone Company Name Message Status Email Draft Created On Add Your Airtable Base ID and Table ID Insert them into the Workflow Configurat
WAutomate YouTube Content Strategy with AI, Apify & Google Sheets
📌 Overview This advanced multi-phase n8n workflow automates the complete research, analysis, and ideation pipeline for a YouTube strategist. It scrapes competitor channels, analyzes top-performing titles and thumbnails, identifies niche trends, gathers audience sentiment from comments, and produces data-driven content ideas—automatically writing them into a structured Google Sheets dashboard. This system is ideal for: YouTube creators Agencies Content strategists Automation engineers Anyone who wants to generate YouTube content ideas backed by real data 🧠 High‑Level Architecture The workflow is split into 5 phases, each building on the previous: Phase 1 — Niche Outliers (Input: User Form) Scrapes 3 high‑quality channels from your niche, extracts their outlier videos, and analyzes why they work. Phase 2 — Broad Niche Insights (Weekly) Scrapes the top trending content in your broad niche (e.g., "AI", "fitness", "personal finance") and logs weekly insights. Phase 3 — Niche Insights (Daily) Scrapes the top videos in your specific micro‑niche daily to keep track of content momentum. Phase 4 — Comment Analysis Analyzes real comments from your channel to understand what your audience li
Create LinkedIn Content with GPT-4 via Telegram Bot & Approval Loop
AI LinkedIn Content Bot with Approval Loop This n8n workflow transforms your Telegram messenger into a personal assistant for creating and publishing LinkedIn posts. You can simply send an idea as a text or voice message, collaboratively edit the AI's suggestion in a chat, and then publish the finished post directly to LinkedIn just by saying "Okay." What You'll Need to Get Started Before you can use this workflow, you'll need a few prerequisites set up. This workflow connects three different services, so you will need API credentials for each: Telegram Bot API Key: You can get this by talking to the "BotFather" on Telegram. It will guide you through creating your new bot and provide you with the API token. New Chat with Telegram BotFather OpenAI API Key: This is required for the "Speech to Text" and "AI Agent" nodes. You'll need an account with OpenAI to generate this key. [OpenAI API Platform](https://platform.openai.com ) Blotato API Key: This service is used to publish the final post to LinkedIn. You'll need a Blotato account and to connect your LinkedIn profile there to get the key. [Blotato platform for social media publishing] Once you have these keys, you can add them to th
SCreate a voice & text Telegram assistant with Lookio RAG and GPT-4.1
Create a Telegram bot that answers questions using Retrieval-Augmented Generation (RAG) powered by Lookio and an LLM agent (GPT-4.1). This template handles both text and voice messages (voice transcribed via a Mistral model by default), routes queries through an agent that can call a Lookio tool to fetch knowledge from your uploaded documents, and returns concise, Telegram-friendly replies. A security switch lets you restrict use to a single Telegram username for private testing, or remove the filter to make the bot public. Who is this for? Internal teams & knowledge workers**: Turn your internal docs into an interactive Telegram assistant for quick knowledge lookups. Support & ops**: Provide on-demand answers from your internal knowledge base without exposing full documentation. Developers & automation engineers**: Use this as a reference for integrating agents, transcription, and RAG inside n8n. No-code builders**: Quickly deploy a chat interface that uses Lookio for accurate, source-backed answers. What it does / What problem does this solve? Provides accurate, source-backed answers: Routes queries to **Lookio so replies are grounded in your documents instead of generic web know
Qualify Real Estate Buyer Leads with GPT-4o & Airtable CRM Integration
🧠 How It Works This AI Agent automatically qualifies property buyer leads from form submissions and sends them directly to your CRM. 🔄 Workflow Steps Form Submission Trigger When a user submits their details via a property inquiry form, the workflow is activated. AI Lead Classification The buyer’s input (budget, location, timeline, etc.) is analyzed by OpenAI. Structured data is extracted, and a lead score (0–100) is generated. Lead Qualification Logic Leads with a score ≥ 70 are marked as qualified. Leads with a lower score can be ignored or stored separately for later review. Follow-Up Actions (for Qualified Leads) An email notification is sent to the real estate agent. A record is created in Airtable to act as a lightweight CRM. ⚙️ How to Set Up 1. Form Setup Replace the default trigger with your preferred source: Typeform, Google Forms, Webflow form, etc. Ensure your form collects the following fields: Name, Email, Budget, Location, Timeline, Property Type 2. Connect Your Credentials Add your OpenAI API key for the LLM node Connect your Gmail account for notifications Link your Airtable base & table to store qualified leads 3. Customize Scoring Logic (Optional) Edit the Infor
WAutomated Customer Onboarding with HubSpot, Gmail, and AI-powered Scheduling
HubSpot Customer Onboarding Automation Overview Streamline your customer onboarding process with this intelligent automation that triggers when new contacts are created in HubSpot. The workflow automatically sends personalized welcome emails, schedules onboarding calls, and assigns customer success managers - ensuring every new customer receives immediate attention and proper support. What This Workflow Does Core Features Real-time Contact Detection**: Monitors HubSpot for new contact creation events via webhooks Intelligent Email Generation**: Uses AI to create personalized welcome emails based on contact information Automated Calendar Scheduling**: Finds available time slots and schedules onboarding calls with new contacts Smart CSM Assignment**: Automatically assigns contacts to appropriate customer success managers Multi-channel Communication**: Sends HTML-formatted emails with professional styling Calendar Integration**: Full Google Calendar management with attendee invitations Business Value Instant Response**: New customers receive immediate welcome communication Personalization at Scale**: Each email is uniquely crafted based on contact details and company information Proce
Automated Blog Post Generator with GPT-4, DALL-E, and Wikipedia for WordPress
WordPress AI Content Creator Overview Transform a few keywords into professionally written, SEO-optimized WordPress blog posts with custom featured images. This workflow leverages AI to research topics, structure content, write engaging articles, and publish them directly to your WordPress site as drafts ready for review. What This Workflow Does Core Features Keyword-to-Article Generation**: Converts simple keywords into comprehensive, well-structured articles Intelligent Content Planning**: Uses AI to create logical chapter structures and content flow Wikipedia Integration**: Researches factual information to ensure content accuracy and depth Multi-Chapter Writing**: Generates coherent, contextually-aware content across multiple sections Custom Image Creation**: Generates relevant featured images using DALL-E based on article content SEO Optimization**: Creates titles, subtitles, and content optimized for search engines WordPress Integration**: Automatically publishes articles as drafts with proper formatting and featured images Business Value Content Scale**: Produce high-quality blog posts in minutes instead of hours Research Efficiency**: Automatically incorporates factual info
C
AI-Powered Multi-Platform Social Media Content Factory with Dynamic System Prompts & GPT-4o
Automated Social Media Content Publishing Factory + System Prompt Composition This workflow is a powerful automation for creating and publishing social media content. It uses AI to generate content and images based on external prompts and schemas, and routes the final posts to various social media platforms. How it Works This workflow operates as a "Content Publishing Factory". It generates platform-optimized content for platforms like LinkedIn, Instagram, Facebook, X (Twitter), Threads, and YouTube Shorts. A key feature is its dynamic system prompt and output schema composition, which pulls content from external Google Docs to ensure consistency. The workflow can also create images to match the content and send the final draft for approval via email before publishing. Use Cases Centralized Content Management**: Store and manage all system prompts and schemas in Google Docs for easy team collaboration and version control. Brand Consistency**: Ensure a consistent brand voice across all social media platforms by centralizing brand guidelines and content rules. Automated Publishing**: Automate content creation, image generation, and publishing, reducing manual effort and design bottle
CWAuto-Generate SEO Product Titles & Descriptions with GPT-4o-mini for Printify
N8N Workflow: Printify Automation - Update Title and Description - AlexK1919 This workflow automates the process of getting products from Printify, generating new titles and descriptions using OpenAI, and updating those products. How it Works This workflow automatically retrieves a list of products from a Printify store, processes them to generate new titles and descriptions based on brand guidelines and custom instructions, and then updates the products on Printify with the new information. It also interacts with Google Sheets to track the status of the products being processed. The workflow can be triggered both manually or by an update in a Google Sheet. Use Cases E-commerce Automation**: Automating content updates for a Printify store. Marketing & SEO**: Generating SEO-friendly or seasonal content for products using AI. Product Management**: Batch-updating product titles and descriptions without manual effort. Setup Instructions Printify API Credentials: Set up httpHeaderAuth credentials for Printify to allow the workflow to get and update products. OpenAI API Credentials: Provide an API key for OpenAI in the openAiApi credentials. Google Sheets Credentials: The workflow requir
A/B Test AI Prompts with Supabase, Langchain Agent & OpenAI GPT-4o
Split Test AI Prompts Using Supabase & Langchain Agent This workflow allows you to A/B test different prompts for an AI chatbot powered by Langchain and OpenAI. It uses Supabase to persist session state and randomly assigns users to either a baseline or alternative prompt, ensuring consistent prompt usage across the conversation. 🧠 Use Case Prompt optimization is crucial for maximizing the performance of AI assistants. This workflow helps you run controlled experiments on different prompt versions, giving you a reliable way to compare performance over time. ⚙️ How It Works When a message is received, the system checks whether the session already exists in the Supabase table. If not, it randomly assigns the session to either the baseline or alternative prompt. The selected prompt is passed into a Langchain Agent using the OpenAI Chat Model. Postgres is used as chat memory for multi-turn conversation support. 🧪 Features Randomized A/B split test per session Supabase database for session persistence Langchain Agent + OpenAI GPT-4o integration PostgreSQL memory for maintaining chat context Fully documented with sticky notes 🛠️ Setup Instructions Create a Supabase table named split_t