HTTP Request workflows
1,200 results — 3 downloadable workflow files, 1,197 source-linked n8n references
WCreate studio-quality car listing photos using Google Drive and Gemini 2.5
Quick overview This workflow manually processes car photos from Google Drive, sends each image plus a studio backdrop and logo to the Google Gemini Image API for background replacement and license-plate branding, and uploads the generated studio-style images back to an output Google Drive folder. How it works Runs when you manually execute the workflow. Lists all image files in a specified Google Drive input folder and processes them one at a time. Downloads the current car photo and also lists and downloads the studio asset images (virtual backdrop and logo) from a separate Google Drive folder. Builds a multimodal request that instructs Google Gemini to classify the shot (exterior vs interior/engine), replace the background accordingly, and swap any visible license plate with the provided logo. Sends the request to the Google Gemini image generation endpoint and receives a generated image in the response. Extracts the generated image, names it using the original filename with a “_studio.jpg” suffix, and uploads it to a specified Google Drive output folder. Setup Add Google Drive OAuth2 credentials and set the input, studio assets, and output folder IDs in the Google Drive nodes’ q
Generate vendor negotiation strategies from pricing data with GPT-4 and SendGrid
Quick overview This workflow ingests vendor pricing via a webhook or a weekly schedule, analyzes variance against historical and market benchmarks, uses OpenAI to draft negotiation strategies for high-priority cases, then emails recommendations through SendGrid and appends the analysis to a Google Sheets tracker. How it works Receives vendor pricing data from a POST webhook endpoint or runs every Monday at 8:00 to process a scheduled input. Normalizes incoming fields (vendor, pricing, benchmarks, volume, currency, renewal date, and notes) into a consistent schema. Calculates price variance versus historical and market benchmarks, estimates potential savings, derives a confidence score, and flags whether the opportunity is high priority. Filters to only high-priority opportunities and waits briefly to help manage API rate limits. Uses OpenAI to generate a structured, under-300-word negotiation strategy including tactics, target and walk-away prices, projected savings range, and risk flags. Formats the AI output into a report, emails it via the SendGrid API, and appends the analyzed record to a Google Sheets tab for tracking. Setup Add an OpenAI API credential for the OpenAI Chat Mod
Generate rich character bibles for writers with GPT-4.1 and Google Sheets
Quick overview This workflow accepts character details via webhook or manual trigger, uses OpenAI to generate psychology, backstory, and emotional arc drafts in parallel, then synthesizes them into a structured “Character Bible” JSON response and optionally appends a log row to Google Sheets. How it works Receives a POST request on an n8n webhook (or runs manually) with character parameters like name, age, archetype, world setting, story role, traits, and constraints. Normalizes the incoming fields, validates the archetype, and derives genre tags from the world setting using Python logic. Sends the enriched character inputs to OpenAI three times in parallel to generate a psychology profile, a backstory, and an emotional arc document. Combines the three AI documents into a single context object and passes it to OpenAI again to synthesize a unified Character Bible with defined sections. Parses the Character Bible into structured JSON fields (identity snapshot, personality core, backstory highlights, emotional arc, voice, and story hooks) and adds metadata like request ID and timestamp. Returns the completed character profile as a JSON webhook response and optionally appends a summary
Audit website SEO with Groq, Google PageSpeed and Google Sheets
Quick overview This workflow audits a submitted website URL by extracting on-page SEO signals from the HTML, pulling Google PageSpeed Insights scores, and using Groq (Llama 3.3 70B) to generate a scored SEO report, then logs the results to Google Sheets and returns the audit via webhook. How it works Receives a POST webhook request containing a JSON body with a target url to audit. Fetches the page HTML for the provided URL and stops with an error response if no HTML is returned. Extracts on-page SEO signals from the HTML (title, meta description/keywords, headings, image alt coverage, links, canonical, robots, viewport, Open Graph/Twitter tags, schema, word count, and language). Calls the Google PageSpeed Insights API for mobile to retrieve performance, accessibility, SEO, and best-practices results. Combines the extracted on-page data with the PageSpeed results and sends it to Groq via a LangChain agent to produce a strict, structured JSON audit with scores, findings, fixes, and top priorities. Parses and validates the AI JSON output, appends/updates the audit record in Google Sheets, and returns the final audit JSON to the webhook caller (or returns a 400 error if parsing fails)
Generate personalized cold email drafts from Google Sheets with Groq and Tavily
Quick overview This scheduled workflow reads leads from Google Sheets, researches each company with Tavily Search, uses Groq (Llama 3.3 70B) to generate a short personalized cold email, and saves the result as a Gmail draft while updating the lead status back in the sheet. How it works Runs on an hourly schedule trigger. Reads lead rows from Google Sheets and keeps only rows that have an email address and are not marked with status "sent". Processes leads one at a time and uses the Tavily Search API to fetch a brief company overview based on the lead’s company name. Sends the lead details plus the Tavily research summary to Groq (Llama 3.3 70B) to generate a JSON-formatted email subject, body, and personalization hook. Parses and validates the AI output to ensure a subject and body are present, and routes failures to an error log. Creates a Gmail draft with the generated subject and body, updates the corresponding row in Google Sheets with the new status and generated fields, and waits briefly before continuing to the next lead. Setup Create and connect a Google Sheets OAuth2 credential, then select the target spreadsheet and sheet in both the Google Sheets read and update steps. C
Rank resumes by job description similarity using Gemini embeddings
Quick overview This workflow collects resume PDFs and a job description via an n8n form, generates embeddings with Google Gemini, calculates cosine similarity between each resume and the job description, and returns a CSV-style ranked list of filenames. How it works Receives resume PDF uploads and a job description text through an n8n Form trigger. Creates a semantic embedding for the job description using the Google Gemini Embeddings API (gemini-embedding-2). Splits the uploaded files into individual resume items and converts each PDF binary to a Base64 string. Sends each Base64-encoded PDF to the Google Gemini Embeddings API to generate a multimodal embedding for the resume. Calculates cosine similarity between the job description embedding and each resume embedding, then sorts resumes from most similar to least similar. Returns a CSV-formatted ranking (Rank, File Name) in the form completion response. Setup Create and add a Google Gemini (PaLM) API credential in n8n with an API key that can access the gemini-embedding-2 model. Ensure the Google Gemini credential is selected on both HTTP requests that create the job description embedding and the resume embeddings. Use the workflo
WScore customer health and churn risk with Salesforce, Gemini, and Gmail
Quick Overview This workflow runs every minute to pull accounts from Salesforce, enrich each customer with order, payment, support, and marketing data via HTTP APIs, analyze the combined profile with Google Gemini, then update the Salesforce account and send Gmail alerts for high-risk customers and execution outcomes. How it works Runs every minute on a schedule and retrieves up to five Salesforce Account records. Processes accounts in batches and initializes a retry counter to control fault-tolerant data fetching. Fetches customer order data via an HTTP API, retrying up to three times with a wait delay and falling back to a backup HTTP endpoint if needed. Fetches payments, support tickets, and marketing engagement for the same customer via HTTP API requests and normalizes everything into a single customer profile. Sends the unified customer profile to Google Gemini to generate a JSON assessment with health score, churn risk, customer type, a summary, and a recommended action. Parses the AI output and emails a high-risk alert via Gmail when the churn risk equals HIGH. Routes by customer type, updates the Salesforce Account description with the AI summary, then logs the result and e
Send multi-carrier shipping quotes via WhatsApp with Gemini and ShipEngine
Quick overview Youtube Video: https://youtu.be/jIGBUW_1Sx0?si=BNXzk95soO0Nefan This workflow responds to incoming WhatsApp messages, validates the sender in HighLevel, uses Google Gemini with Redis memory to collect shipment details, fetches multi-carrier rates from ShipEngine, applies a commission markup, logs the quote to Google Sheets, and sends the best options back via WhatsApp. How it works Triggers when a new WhatsApp message is received. Looks up the sender’s phone number in HighLevel and stops outreach if the contact is marked Do Not Disturb. Uses a Google Gemini-powered agent with Redis chat memory to collect and confirm shipment details, creating the contact in HighLevel if needed or marking them DND if they opt out. If shipment data collection is complete, requests shipping rates from the ShipEngine Rates API for the configured carriers. Calculates a commission markup, filters for valid rates, and selects the cheapest and fastest options to format a quote message. Appends the quote text and metadata to a Google Sheets spreadsheet and sends the quote back to the customer via WhatsApp. Setup Connect WhatsApp Business credentials for both the WhatsApp trigger and send acti
WAnswer HR and IT policy questions on Telegram with OpenAI and Postgres PGVector
Quick overview This workflow builds a searchable HR policy knowledge base from a PDF and runs a Telegram helpdesk bot that answers employee questions from that content, handling both text messages and voice notes via OpenAI transcription and retrieval from a Postgres PGVector vector store. How it works Manually runs once to download an HR policy PDF over HTTP and extract its text. Splits the extracted policy text into chunks, generates OpenAI embeddings, and inserts the vectors into a Postgres PGVector store. Triggers on new Telegram messages sent to your bot. Routes incoming Telegram updates by type, sending text directly to the assistant, transcribing voice messages with OpenAI, or replying that other formats are unsupported. Uses an OpenAI chat model with Postgres-backed chat memory and a PGVector retrieval tool to answer questions grounded in the stored policy content. Sends the final answer back to the user in Telegram. Setup Add an OpenAI API key for embeddings, chat completions, and audio transcription. Set up a Postgres database with the PGVector extension enabled and add the Postgres credentials for both the vector store and chat memory. Create a Telegram bot with BotFathe
Send SAP Business One low stock alerts via WhatsApp and email
Quick overview This workflow polls SAP Business One every 6 hours for active, unfrozen items whose on-hand quantity is below the minimum inventory level, then sends de-duplicated low-stock alerts via WhatsApp (Meta API) and email, storing the last-alerted quantity in an n8n Data Table to prevent repeated notifications. How it works Runs every 6 hours on a schedule. Logs in to SAP Business One Service Layer and queries the Items endpoint for valid, non-frozen items where QuantityOnStock is less than MinInventoryQuantity. Splits the SAP response into one item per record and skips the alert flow if no items are returned. Looks up each item’s previous alert state in an n8n Data Table and only continues if the item has never been alerted before or its stock has dropped further since the last alert. Formats a WhatsApp message and an HTML email (including an “out of stock” message when QuantityOnStock is 0). Sends the alert via the WhatsApp Business Cloud API (Meta Graph API) and via SMTP email, then upserts the latest alerted quantity to the Data Table. Logs out of SAP Business One. Setup Configure SAP Business One Service Layer access by updating the server URL in the SAP login, item qu
Publish scheduled LinkedIn posts from Twenty CRM via Unipile and SharePoint
Quick overview Automate your LinkedIn content calendar. This workflow fetches scheduled posts from a PostgreSQL database (Twenty CRM), downloads attached media from SharePoint, and publishes them seamlessly to LinkedIn using Unipile. How it works A Schedule Trigger (Cron) fires every 5 minutes to check for pending social media posts A PostgreSQL node queries the database to fetch posts marked as 'SCHEDULED' where the scheduled time has passed. A Split In Batches node loops through the fetched posts one by one to prevent API concurrency issues. An IF node checks if there is a SharePoint attachment URL included in the database record. If an attachment exists, Microsoft Graph API nodes resolve the SharePoint Drive Item and download the image as a binary file HTTP Request nodes send the text (and binary attachment, if any) to Unipile's API to natively publish the LinkedIn post. A final PostgreSQL node updates the CRM record to 'POSTED' and saves the live LinkedIn post URL back to the database. Setup Add your PostgreSQL credentials and verify that the SQL query matches your Twenty CRM table and schema names. Connect your Microsoft Entra ID (Azure AD) account to authenticate SharePoint a
Alert on API contract drift using Data Tables and Slack
Quick overview This workflow runs on a schedule, polls a JSON API endpoint, derives and stores a snapshot of the response schema in an n8n Data Table, and posts a Slack alert only when the response structure or field types change. How it works Runs every 6 hours on a schedule. Loads the configured API endpoint URL, an endpoint key used for snapshotting, and the target Slack channel. Fetches the endpoint with retries and skips the run if the response is not a valid JSON object or array. Derives a normalized schema (field paths, types, required/optional, and nullability) and compares it to the previous schema stored in an n8n Data Table. Upserts the latest schema snapshot back into the n8n Data Table under the same endpoint key. Posts a formatted drift report to Slack only when a high- or medium-severity contract change is detected. Setup Create an n8n Data Table (for example, “API Contract Snapshots”) with text columns named endpointKey and schema_object, and select its ID in both the read and upsert steps. Add a Slack API credential in n8n. Update the Settings values for endpointUrl, endpointKey, and slackChannel (Slack channel ID) to match your API and destination channel. Run the
Create social posts from RSS feeds with OpenAI, Unsplash, and Buffer
Quick overview This workflow polls an RSS feed for new articles, uses OpenAI to generate platform-specific copy for X, LinkedIn, and Instagram, pulls a related random image from the Unsplash API, and creates queued posts in Buffer for each social channel. How it works Runs every hour and reads the latest items from the configured RSS feed. Sends each article title and snippet to OpenAI to generate three ready-to-post messages for X, LinkedIn, and Instagram. Requests a random related image from the Unsplash API using the article title as the search query. Creates an image post in Buffer for X using the generated Twitter text and the Unsplash image URL. Creates an image post in Buffer for LinkedIn using the generated LinkedIn text and the Unsplash image URL. Creates an image post in Buffer for Instagram using the generated Instagram caption and the Unsplash image URL. Setup Add an OpenAI API credential in n8n (or replace the included OpenAI credential) for the OpenAI node. Create a Buffer account, connect your X/Twitter, LinkedIn, and Instagram profiles, and add your Buffer API credential in n8n. Create an Unsplash developer app and replace the client_id value in the HTTP Request URL
Rank resumes by job description similarity using OpenRouter embeddings
Quick overview This workflow collects resume PDFs and a job description through an n8n form, extracts text (with OCR for scanned PDFs), generates embeddings via OpenRouter, and ranks the resumes by cosine similarity to the job description, returning a CSV-style list of ranked filenames. How it works Receives resume PDFs and a job description from an n8n form submission. Creates an embedding for the job description using OpenRouter’s embeddings endpoint with openai/text-embedding-3-large. Processes each uploaded resume one at a time, converting the PDF to text using built-in PDF extraction. If the extracted text is too short (likely scanned), sends the PDF to OpenRouter chat completions with the file-parser plugin (mistral-ocr) to OCR the document and return plain text. Creates an embedding for each resume’s extracted text (from direct extraction or OCR) using openai/text-embedding-3-large. Computes cosine similarity between each resume embedding and the job description embedding, then sorts resumes from most similar to least similar. Returns a completion page containing a CSV-style list of rank and filename. Setup Add an OpenRouter API credential in n8n and select it on the embeddi
Monitor n8n security advisories from GitHub, NVD, JPCERT and JVN with Claude and Slack
Quick overview This workflow runs hourly to monitor n8n-related security information from GitHub Advisories, NVD (NIST) CVE data, GitHub n8n security release notes, and JPCERT/JVN RSS feeds, then summarizes new findings with Anthropic Claude and posts an alert to Slack. How it works Runs every hour on a schedule. Fetches the latest items from the GitHub Advisory API, the NVD CVE API, the n8n GitHub Releases Atom feed, and the JPCERT and JVN RSS feeds. Parses the XML feeds, normalizes all sources into a single list of n8n-related security items, and removes duplicates and previously notified items using stored IDs. Checks whether any new items were found and stops if there are none. Sends the new items to Anthropic Claude to generate a concise Japanese summary including importance, affected product, and urgency. Formats a Slack Block Kit message and posts the alert to the configured Slack channel. Saves the notified item IDs to workflow static data to prevent repeat notifications in future runs. Setup Create an NVD API key and add it as an HTTP Header Auth credential named (or mapped to) NVD_API_Key with header name apiKey. Add an Anthropic API credential for the Claude model used i
Create blog drafts, images, and LinkedIn posts with Tavily, Gemini, and OpenAI
Quick overview This workflow receives a blog request via webhook, researches the topic with Tavily, generates a long-form HTML article using Google Gemini, creates two images via the kie.ai API, stores assets in Google Drive, updates a Google Sheets tracker, and notifies a Slack user with links. How it works Receives a POST webhook request containing a Google Sheets row payload with an “AI Blog Title” and related metadata. Searches the web for supporting sources using Tavily and passes the research plus the title to Google Gemini to generate the full article as HTML. Converts the generated content to HTML, creates a temporary HTML file in Google Drive, downloads it to set the correct text/html MIME type, re-uploads it as the final Drive file, and deletes the temporary file. Sends the finished article to OpenAI to generate a short LinkedIn post, then updates the matching row in Google Sheets with a Done flag, the article folder/link value, and the LinkedIn post text. Creates a cover image and an in-article infographic by submitting two text-to-image jobs to the kie.ai API, polling until each job succeeds, downloading each resulting image, and uploading both images to Google Drive. W
Create social posts from RSS with OpenAI, LinkedIn, X, and Bluesky
Quick overview This workflow runs daily, reads the Hacker News frontpage RSS feed, uses OpenAI to extract key insights and ghostwrite social posts, then publishes the generated content to LinkedIn, X, and Bluesky. How it works Runs every day at 08:00 using a schedule trigger. Fetches the latest items from the Hacker News frontpage RSS feed and keeps only the first three articles. Prepares a clean title, content snippet, and link for each article. Sends each article to OpenAI Chat Completions (gpt-4o-mini) to extract three non-obvious insights. Sends the insights to OpenAI Chat Completions again to generate platform-specific posts (LinkedIn, X, and Bluesky) in a defined voice and JSON format. Parses and clamps the generated text to length limits and then posts to LinkedIn and X while also authenticating to Bluesky and creating a Bluesky post. Setup Add an OpenAI API credential for the two OpenAI Chat Completions HTTP requests. Add LinkedIn and X (Twitter) credentials and confirm the target posting identity/settings for each platform. Replace the Bluesky handle and app password in the Bluesky login request and ensure Bluesky app passwords are enabled for the account. (Optional) Updat
CScrape Tunisia IT job listings with Groq, Firecrawl, Airtable, and Resend
Quick overview This workflow runs daily or via chat to scrape Tunisia-focused IT job listings from predefined company career pages using Firecrawl, saves results into Airtable tables, and emails a daily digest through Resend, with job parsing and orchestration handled by a Groq-powered AI agent. How it works Triggers either on a daily 7AM schedule or when an n8n Chat message is received. On scheduled runs, rotates through a fixed list of companies, builds a prompt for the chosen company (including pagination rules), and sets the Airtable table name. Uses a Groq (Llama 3.3 70B) AI agent to decide which allowed career-page URLs to scrape and how to extract job listings into structured fields. Scrapes the requested career pages via the Firecrawl API and parses the returned main content into jobs with title, company, location, description, and URL. When saving is requested, uses Airtable’s API to check for the target table, create it if missing, and insert the job records. For scheduled runs, emails the AI-generated job digest to the configured recipient using the Resend API; otherwise it outputs the response for the chat session. Setup Add credentials and values for Firecrawl (API key
Run a multi-agent research and publishing pipeline with GPT-4o, Tavily and Notion
Quick overview Three specialized AI agents: Researcher, Writer, and Reviewer, collaborate autonomously to research a topic, draft content, and quality-check it through a self-correcting review loop with a circuit breaker. Approved content is published directly to Notion with full audit metadata. How it works A form submission captures the research topic, content format (blog post, exec summary, or LinkedIn post), target audience, and maximum revision cycles allowed before the circuit breaker fires. The Researcher Agent uses GPT-4o and Tavily Search to gather authoritative sources, extract key facts and contradictions, and produce a validated research brief. Search depth and result count are automatically scaled to the requested format. The Writer Agent receives the research brief and produces a structured draft in the correct format. On revision cycles it also receives the Reviewer's specific rejection notes and must address them directly. A Circuit Breaker sits between the Writer and Reviewer. If the number of Reviewer rejections reaches the user-defined maximum, the draft is force-approved and the pipeline proceeds to publishing without another LLM call. The Reviewer Agent evalua
W
Reply to Instagram comments with Groq, Telegram approval, and Google Sheets logging
Quick Overview This workflow monitors new Instagram comments, generates a draft reply with Groq (OpenAI-compatible chat completions), sends it to Telegram for YES/NO approval, and logs the decision in Google Sheets before optionally posting the approved reply back to Instagram. How it works Runs every 10 minutes, reads existing comment IDs from Google Sheets, then fetches your 10 most recent Instagram media items and their comments via the Instagram Graph API. Filters out comments made by your own Instagram account and limits processing to recent comments. Skips any comment already logged in Google Sheets, and sends new comments (with the post caption as context) to Groq to generate a short draft reply. Sends the draft reply to a Telegram chat for approval and appends a “pending_review” row to Google Sheets with the Telegram message ID for tracking. Runs every 3 minutes, polls the Telegram Bot API for new updates, and keeps an offset so only new messages are processed. For Telegram messages that are replies to an approval request, matches the replied-to message ID to the pending row in Google Sheets and routes the decision. If the reviewer replies YES, posts the draft as an Instagr
Sync Etsy customers with Meta custom audiences using hashed emails
Quick overview This workflow runs daily to fetch recent Etsy receipt emails via an Etsy Shop Operations sub-workflow, normalizes and SHA-256 hashes the emails, and either previews the hashed payload or uploads it to a Meta Custom Audience using the Meta Graph API. How it works Runs every day on a schedule. Sets sync parameters (lookback window, maximum receipts, dry-run mode, Meta audience ID, and API version). Requests recent Etsy receipt records from a connected Etsy Shop Operations workflow and extracts buyer/payment email addresses. Normalizes, deduplicates, and SHA-256 hashes the email addresses to build a Meta Custom Audience upload payload. If dry-run is enabled, returns counts and a small sample of hashes for verification. If dry-run is disabled, sends the hashed EMAIL payload to the Meta Graph API endpoint for the specified Custom Audience. Setup Configure and connect the “Get Etsy customer receipts” execute-workflow step to an Etsy Shop Operations (PRO) workflow that returns receipt records including buyer/payment emails. Create or choose a Meta Custom Audience and paste its ID into the configuration values. Add a Meta access token to the HTTP header auth for the upload r
Manage Etsy OAuth 2.0 tokens and API headers between workflows
Quick Overview This workflow manages Etsy OAuth 2.0 (PKCE) authorization in n8n, stores access and refresh tokens in workflow static data, refreshes tokens when needed, and returns ready-to-use Etsy API request headers to other workflows via an Execute Workflow call. How it works Runs either manually to start authorization, via a webhook callback from Etsy after consent, or when another workflow requests a token. Loads the Etsy app configuration (client ID, shared secret, redirect URI, scopes) and routes the run to start authorization, handle the callback, or return an access token. When starting authorization, generates a PKCE verifier/challenge and returns an Etsy authorization URL while storing the pending state for validation. When Etsy redirects to the callback webhook, validates the returned state and authorization code and then exchanges the code for tokens using the Etsy OAuth token endpoint. Stores the authorized access and refresh tokens (with expiry timestamps) in n8n workflow static data and returns a confirmation response to the callback request. When another workflow requests a token, loads the stored refresh token, refreshes the access token with Etsy if it is missin
WCapture and log email trade instructions with Gmail, OpenAI, and Google Sheets
Quick overview This workflow polls Gmail for emails with the subject “Trade Instruction”, uses OpenAI Chat Completions to extract trade fields from the email body, logs valid and failed extractions to separate tabs in Google Sheets, and sends confirmation and ops alert emails via Gmail. How it works Triggers every minute by polling Gmail for new messages matching the query subject:"Trade Instruction". Extracts the email body, sender address, subject, and message/thread IDs and generates a trade reference and timestamps. Sends the email body to OpenAI (Chat Completions) to return a JSON object containing asset, ticker, action, quantity, price, client, and notes. Parses the returned JSON, validates that asset is present and quantity is greater than zero, and marks the result as valid or failed with a reason. Appends valid trade instructions to the “Valid Trades” tab in Google Sheets, or appends failures (including a snippet and error reason) to the “Failed Trades” tab. For valid trades, replies to the original email in Gmail with a receipt confirmation and emails the ops team a capture notification, and for failed trades, emails the ops team a manual-review alert with the original em
Create SEO blog posts from YouTube videos with Gemini and Google Docs
Quick overview Submit a YouTube URL and get a publish-ready SEO blog post. Three Gemini agents research the video, write the article, and score it. Articles scoring 7+ auto-publish to Google Docs. Lower scores route to Telegram for review. How it works The workflow starts when a POST request hits the webhook with a YouTube URL, optional target keyword, and tone. A Code node extracts the video ID and validates the URL format, then an HTTP Request node calls YouTube Data API v3 to pull the video title, description, channel name, and tags. Three Basic LLM Chain nodes run in sequence, each powered by a Google Gemini Flash sub-node connected via the native ai_languageModel port. The first chain (Research Agent) receives the video context and returns a structured JSON brief containing an SEO title, four H2 headings, primary keyword, secondary keywords, key insights, target audience, and a content angle. A Code node parses this output with a JSON fallback in case Gemini adds unexpected formatting. The second chain (Writer Agent) takes the research brief and writes a complete 1,000–1,400 word blog post following strict rules: flowing paragraphs over bullet lists, primary keyword used 3–5 t