Harsh Maniya This workflow corresponds to n8n.io template #6926 — we link there as the canonical source.
This workflow follows the Chainllm → Anthropic Chat recipe pattern — see all workflows that pair these two integrations.
The workflow JSON
Copy or download the full n8n JSON below. Paste it into a new n8n workflow, add your credentials, activate. Full import guide →
{
"id": "KxLD27K9btIxsRZo",
"meta": {
"templateCredsSetupCompleted": true
},
"name": "Whatsapp Deep research",
"tags": [],
"nodes": [
{
"id": "f3a0dc1a-f64b-4f2e-945a-2cfdfe607e45",
"name": "Fetch Whatsapp Request",
"type": "n8n-nodes-base.webhook",
"position": [
-528,
0
],
"parameters": {
"path": "fetch-whatsapp-request",
"options": {},
"httpMethod": "POST"
},
"typeVersion": 2
},
{
"id": "51fc09d1-55e2-4198-977b-79066559d246",
"name": "Perform Deep Research",
"type": "n8n-nodes-base.perplexity",
"position": [
-224,
0
],
"parameters": {
"model": "sonar-pro",
"options": {},
"messages": {
"message": [
{
"content": "=You are a specialized AI Research Analyst. Your task is to perform a deep, comprehensive, and objective research analysis on the user-provided topic using your advanced sonar search capabilities. The final output must be a well-structured report formatted in Markdown, suitable for delivery via an API.\n\n**User Query:**\n\"{{ $json.body.Body }}\"\n\n**Required Output Structure:**\n\n1. **Title:** A clear and concise title for the research report.\n2. **Overview (1-2 paragraphs):** A high-level executive summary of the topic. This should give the user the most important information at a glance.\n3. **Key Points (Bulleted List):** A list of 3-5 of the most critical facts, findings, or aspects of the topic.\n4. **Detailed Analysis:** A more in-depth exploration of the topic. Break this section down with subheadings (using bold text) as needed. Cover background, current status, different perspectives, and relevant data.\n5. **Conclusion:** A brief concluding paragraph that summarizes the main takeaways or future outlook.\n6. **Sources:** A numbered list of the primary sources used for this analysis.\n\n**Guidelines:**\n- **Tone:** Maintain a neutral, objective, and factual tone.\n- **Accuracy:** Prioritize information from reputable sources. Synthesize findings from multiple sources to provide a balanced view.\n- **Formatting:** Strictly adhere to the Markdown structure specified above. Use `*bold*` for headings and bullet points (`-` or `*`) for lists.\n- **Clarity:** Write in clear, concise language. Avoid overly technical jargon where possible, or explain it if necessary.\n- **Completeness:** The generated report should be a self-contained document based solely on the provided user query. Do not ask follow-up questions."
}
]
},
"simplify": true,
"requestOptions": {}
},
"credentials": {
"perplexityApi": {
"name": "<your credential>"
}
},
"typeVersion": 1
},
{
"id": "783daa3b-db9e-4a8c-b0e1-60679ee293cf",
"name": "Refine Generated Output for Whatsapp",
"type": "@n8n/n8n-nodes-langchain.chainLlm",
"position": [
112,
0
],
"parameters": {
"text": "=You are an expert content editor who transforms raw text into perfectly formatted summaries for WhatsApp. Your task is to take the provided research report and polish it into a message that is readable, engaging, and renders correctly on WhatsApp.\n\n**Raw Research Report:**\n{{ $json.message }}\n\n**Your Instructions:**\n\n1. **Preserve Core Content:** Do not change the facts, data, or the meaning of the original text. Your job is to edit for style and formatting only.\n\n2. **Add Emojis:**\n * Add a single, relevant emoji at the beginning of the Title.\n * Add relevant emojis to the main section headers and to each bullet point in the \"Key Points\" list.\n\n3. **Strict Formatting Rules for WhatsApp:**\n * **Bolding:** Apply bold formatting by enclosing headers in **single asterisks**. For example: `*Overview:*`. Do NOT use double asterisks.\n * **Bullet Points:** Use a standard hyphen (`-`) followed by a single space for each bullet point.\n * **Spacing:** Ensure there is a **double line break** (press Enter twice) between each major section to create clean visual separation.\n * **Citations:** Preserve the numerical inline citations like `[1]`, `[2]`, etc., exactly as they are.\n\n4. **Character Limit:** The entire final output MUST be less than 1600 characters to comply with WhatsApp's message limit. Be concise.\n\n5. **No Preamble:** Your output must begin *directly* with the report's title. Do not include any introductory phrases like \"Here is the polished version:\".\n\n**Example of Desired Final Output Format:**\n\n*Example:*\n\ud83e\udd16 *Title:* Example Report\n\ud83d\udcdc *Overview:*\nThis is a summary of the topic. It is clear and concise.\n\n\ud83c\udfaf *Key Points:*\n- \u2705 This is the first key point.\n- \u2699\ufe0f This is the second key point.\n\n---\n\nProduce the final, polished WhatsApp-ready report now, strictly following all the rules and the format shown in the example.\n\n**Remember whatsapp allows only 1600 characters to send via message so your output must be less than or equal to that.**",
"batching": {},
"promptType": "define"
},
"typeVersion": 1.7
},
{
"id": "e613cdb5-7ba4-4519-b15e-a0174fdb55d6",
"name": "Send Output in Whatsapp",
"type": "n8n-nodes-base.twilio",
"position": [
640,
0
],
"parameters": {
"to": "={{ $('Fetch Whatsapp Request').item.json.body.From.replace('whatsapp:', '') }}",
"from": "={{ $('Fetch Whatsapp Request').item.json.body.To.replace('whatsapp:', '') }}",
"message": "={{ $json.text }}",
"options": {},
"toWhatsapp": true
},
"credentials": {
"twilioApi": {
"name": "<your credential>"
}
},
"typeVersion": 1
},
{
"id": "ede2a3ae-6dc9-486e-999f-dcb9737fd79b",
"name": "Claude Sonnet 4",
"type": "@n8n/n8n-nodes-langchain.lmChatAnthropic",
"position": [
112,
208
],
"parameters": {
"model": {
"__rl": true,
"mode": "list",
"value": "claude-sonnet-4-20250514",
"cachedResultName": "Claude 4 Sonnet"
},
"options": {}
},
"credentials": {
"anthropicApi": {
"name": "<your credential>"
}
},
"typeVersion": 1.3
},
{
"id": "3c93f1c8-e00e-4791-b10f-02f9c0d82630",
"name": "Sticky Note",
"type": "n8n-nodes-base.stickyNote",
"position": [
-624,
-224
],
"parameters": {
"width": 288,
"height": 560,
"content": "### \ud83d\udcf2 WhatsApp Gateway\n\n**Function:** This is the front door. It catches incoming messages from a user via the Twilio webhook.\n\n**Trigger:** Receives a POST request from Twilio.\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n**Output:** Passes the user's message (`body.Body`), their phone number (`body.From`), and the bot's number (`body.To`) to the next nodes.\n"
},
"typeVersion": 1
},
{
"id": "198ba40c-5f71-4cef-a69e-fb621829281d",
"name": "Sticky Note1",
"type": "n8n-nodes-base.stickyNote",
"position": [
-320,
-224
],
"parameters": {
"color": 7,
"width": 288,
"height": 560,
"content": "### \ud83e\udde0 The AI Research Engine\n\n**Function:** This node is the core researcher. It takes the user's simple question and performs a deep, multi-source dive using Perplexity's powerful `Sonar Pro` model.\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n* **Input:** The raw text query from the user.\n* **Output:** A detailed, structured report in raw Markdown, complete with citations and sources. This is the \"first draft\"."
},
"typeVersion": 1
},
{
"id": "eada649f-b85c-4696-a670-eeafb6ac2773",
"name": "Sticky Note2",
"type": "n8n-nodes-base.stickyNote",
"position": [
-16,
-224
],
"parameters": {
"color": 3,
"width": 528,
"height": 560,
"content": "### \ud83c\udfa8 The Content Stylist\n\n**Function:** This node acts as an expert editor. It takes the raw research from Perplexity and polishes it into a perfect WhatsApp message.\n\n**Key Actions:** Adds emojis, enforces correct formatting (`bold`), shortens the text to fit the character limit, and removes all conversational filler.\n* **Input:** The raw Markdown report.\n* **Output:** A single, polished, and user-ready text string.\n"
},
"typeVersion": 1
},
{
"id": "097b9242-ce57-48a2-9be5-f3e59298b634",
"name": "Sticky Note3",
"type": "n8n-nodes-base.stickyNote",
"position": [
528,
-224
],
"parameters": {
"color": 6,
"width": 320,
"height": 560,
"content": "### \ud83d\ude80 WhatsApp Dispatch\n\n**Function:** This is the final step. It sends the beautifully formatted research summary back to the user who asked for it.\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n* **Clever Logic:** It automatically swaps the `From` and `To` numbers from the first node to ensure the reply always goes to the right person.\n* **Action:** Sends the final text from the \"Content Stylist\" node via the Twilio API. Mission complete!"
},
"typeVersion": 1
},
{
"id": "c5462997-d89d-4ae6-9d62-d5bca26bd02f",
"name": "Sticky Note4",
"type": "n8n-nodes-base.stickyNote",
"position": [
-1008,
-224
],
"parameters": {
"color": 5,
"width": 368,
"height": 560,
"content": "## \u2b50 Workflow Summary\n\n> ### \ud83e\udd16 WhatsApp AI Research Assistant\n>\n> This workflow powers a fully automated, on-demand research assistant that operates directly within WhatsApp. It's designed to take any user query received via **Twilio**, understand its core topic, and deliver a comprehensive yet beautifully formatted summary directly back to the user's chat.\n>\n> The process is a seamless four-step data journey:\n> 1. **Catch (\ud83d\udcf2):** The workflow activates when a user's message is received by the webhook.\n> 2. **Research (\ud83e\udde0):** The raw query is immediately sent to **Perplexity Sonar** to perform a deep, multi-source web search, generating a fact-based, structured report.\n> 3. **Polish (\ud83c\udfa8):** That raw report is then passed to a **Claude** model, which acts as an expert content editor. It shortens the text, adds engaging emojis, and formats it perfectly for a mobile screen.\n> 4. **Reply (\ud83d\udcac):** The final, polished answer is instantly dispatched back to the original user through **Twilio**, completing the entire request in seconds."
},
"typeVersion": 1
}
],
"active": false,
"settings": {
"executionOrder": "v1"
},
"versionId": "ffca0cbc-b83b-4419-910d-4c05e17be6bf",
"connections": {
"Claude Sonnet 4": {
"ai_languageModel": [
[
{
"node": "Refine Generated Output for Whatsapp",
"type": "ai_languageModel",
"index": 0
}
]
]
},
"Perform Deep Research": {
"main": [
[
{
"node": "Refine Generated Output for Whatsapp",
"type": "main",
"index": 0
}
]
]
},
"Fetch Whatsapp Request": {
"main": [
[
{
"node": "Perform Deep Research",
"type": "main",
"index": 0
}
]
]
},
"Refine Generated Output for Whatsapp": {
"main": [
[
{
"node": "Send Output in Whatsapp",
"type": "main",
"index": 0
}
]
]
}
}
}
Credentials you'll need
Each integration node will prompt for credentials when you import. We strip credential IDs before publishing — you'll add your own.
anthropicApiperplexityApitwilioApi
Pro
For the full experience including quality scoring and batch install features for each workflow upgrade to Pro
About this workflow
Ever wished you could get a deep, multi-source research report on any topic, delivered directly to your WhatsApp chat in seconds? This workflow transforms your WhatsApp into a powerful, on-demand research assistant, perfect for students, professionals, and curious minds.
Source: https://n8n.io/workflows/6926/ — original creator credit. Request a take-down →
Related workflows
Workflows that share integrations, category, or trigger type with this one. All free to copy and import.
This template uses the Apify LinkedIn Profile Scraper, which is a community node only available in self-hosted n8n installations. The LinkedIn scraping step is optional and can be removed for n8n Clou
HTTP Request, Perplexity, Slack +8
Automatically reads every reply to your cold email campaigns in Instantly.ai, uses Claude AI to understand the intent, and takes the right action . No need ofmanual inbox checking needed. A lead repli
Chain Llm, Anthropic Chat, Slack +3
Requirements: GitHub API token ( scope), Anthropic API key (Claude Sonnet 4.5), Slack Bot Token (optional)
Form Trigger, HTTP Request, Chain Llm +2
Automatically score product usage signals from Amplitude cohorts and route hot leads to sales with enriched context.
HTTP Request, Perplexity, Agent +4
This workflow is a powerful, fully automated web query and semantic reranking system that allows users to perform precise, detailed searches, intelligently rank search results and provide high-quality
Output Parser Autofixing, Output Parser Structured, Chain Llm +4