How to Automate Daily AI Research with n8n and DeepSeek: Zero-Code Guide

TL;DR

If you're searching for "how to automate AI research with n8n," this tutorial gets you a working daily AI news digest pipeline in 30 minutes — no coding required. You'll connect n8n's Schedule Trigger to an HTTP Request node that pulls headlines from NewsAPI, pipe them through DeepSeek's API for intelligent summarization and categorization, and deliver the final digest to your Slack or email. Everything is drag-and-drop.

What You'll Learn

• Set up a self-hosted n8n instance in 5 minutes with Docker
• Configure the Schedule Trigger for daily automated runs
• Use the HTTP Request node to fetch real-time news from any REST API
• Call DeepSeek's chat API inside n8n to summarize, categorize, and extract key insights
• Format the AI output and deliver it to Slack, email, or Notion
• Debug common n8n workflow errors (JSON parsing, rate limits, empty responses)

Prerequisites

💡 No DeepSeek API key yet? Get one at platform.deepseek.com — new users get free credits. You can also substitute OpenAI, Claude, or any OpenAI-compatible endpoint.

Step 1: Spin Up n8n with Docker (5 Minutes)

The fastest way to get n8n running locally:

docker run -d \
  --name n8n \
  -p 5678:5678 \
  -v n8n_data:/home/node/.n8n \
  -e N8N_SECURE_COOKIE=false \
  docker.n8n.io/n8nio/n8n

Once the container starts, open http://localhost:5678 in your browser. You'll see the n8n editor — a blank canvas where workflows are built by connecting nodes.

Quick setup steps:

1. Create an owner account (email + password)
2. Skip the onboarding survey
3. You're now on the workflow editor — ready to build

Troubleshooting: If port 5678 is already in use, change -p 5678:5678 to -p 5679:5678 and visit http://localhost:5679.

Step 2: Add a Schedule Trigger

Every automation starts with a trigger. For a daily digest, use the Schedule Trigger node.

1. Click the + button on the canvas or press Tab
2. Search for "Schedule" and select Schedule Trigger
3. In the node settings panel, configure:
   • Trigger Interval: Days
   • Days Between Triggers: 1
   • Trigger at Hour: 8 (runs at 8:00 AM)
   • Trigger at Minute: 0

Click Execute Node to test — you'll see a timestamp output confirming the trigger works. The node produces a single JSON object with the current date and time.

Step 3: Fetch News with the HTTP Request Node

Now, pull real news headlines. Add an HTTP Request node:

1. Click the + icon connected to the Schedule Trigger
2. Search for "HTTP Request" and select it
3. Configure the node:

{
  "Method": "GET",
  "URL": "https://newsapi.org/v2/top-headlines",
  "Query Parameters": {
    "country": "us",
    "category": "technology",
    "apiKey": "YOUR_NEWSAPI_KEY",
    "pageSize": 10
  },
  "Options": {
    "Response Format": "JSON"
  }
}

What this does: fetches the top 10 US technology headlines from NewsAPI. The response is a JSON object containing an articles array, where each article has title, description, url, and source fields.

Click Execute Node to test. You should see a JSON response with article data.

🔄 Alternative data sources: Replace the URL with any RSS feed (https://feeds.feedburner.com/TechCrunch), Reddit JSON API (https://www.reddit.com/r/MachineLearning/hot.json), or Hacker News API (https://hacker-news.firebaseio.com/v0/topstories.json). The HTTP Request node works with any REST endpoint.

Step 4: Summarize with the AI Node (DeepSeek)

This is where the magic happens. n8n has a built-in AI node group with an OpenAI Chat Model node that works with any OpenAI-compatible API — including DeepSeek.

4a. Add OpenAI Chat Model credentials

1. Add an OpenAI Chat Model node (found under "AI" in the node panel)
2. Click the + next to "Credential for OpenAI API"
3. Fill in:
   • API Key: your DeepSeek API key
   • Base URL: https://api.deepseek.com/v1 ⚠️ Important: don't forget /v1

4b. Chain nodes: HTTP Request → AI Chain

Don't connect HTTP Request directly to the AI node. Instead, use n8n's AI Chain pattern:

1. Add a Basic LLM Chain node (search "Basic LLM")
2. Connect: HTTP Request → Basic LLM Chain
3. Add OpenAI Chat Model as a sub-node inside Basic LLM Chain

4c. Craft the prompt

In the Basic LLM Chain node, set:

User Prompt (use the expression editor to pull data from the HTTP Request node):

Analyze these technology news headlines and create a concise daily digest.
For each article, extract:
1. A one-sentence summary
2. The key company/organization mentioned
3. A relevance score (1-5)

Then categorize the entire digest into 2-3 themes and write a 2-sentence overall trend analysis.

Here are the articles:
{{ $json.articles }}

💡 The {{ $json.articles }} expression pulls the articles array from the HTTP Request output. n8n's expression system references each node's output with $json.

System Prompt:

4d. Model settings

In the OpenAI Chat Model sub-node:

• Model: deepseek-chat
• Temperature: 0.3 (lower = more factual)
• Max Tokens: 2000

Click Execute Node to test the full chain. You should see DeepSeek's response — a structured digest of the fetched articles.

Step 5: Format and Deliver the Digest

The AI output is raw text inside n8n. Clean it up and send it somewhere useful.

Option A: Send to Slack

1. Add a Slack node (search "Slack")
2. Connect from Basic LLM Chain → Slack
3. Configure:
   • Resource: Message
   • Operation: Post
   • Channel: #ai-research (or your channel name)
   • Text: {{ $json.text }} (the full AI digest)

Slack credential setup: Create a Slack app at api.slack.com, add the chat:write OAuth scope, and paste the Bot Token into n8n.

Option B: Send via Email

1. Add an Email (SMTP) node
2. Configure:
   • From Email: digest@yourdomain.com
   • To Email: team@yourdomain.com
   • Subject: 🤖 Daily AI Research Digest — {{ $now.format('YYYY-MM-DD') }}
   • Body: {{ $json.text }}

Option C: Save to Notion (Advanced)

For a searchable knowledge base, pipe the digest into Notion:

1. Add another HTTP Request node
2. Configure to call the Notion API:

{
  "Method": "POST",
  "URL": "https://api.notion.com/v1/pages",
  "Headers": {
    "Authorization": "Bearer YOUR_NOTION_TOKEN",
    "Notion-Version": "2022-06-28"
  },
  "Body": {
    "parent": { "database_id": "YOUR_DB_ID" },
    "properties": {
      "Name": { "title": [{ "text": { "content": "Daily Digest" } }] },
      "Date": { "date": { "start": "{{ $now.format('YYYY-MM-DD') }}" } }
    },
    "children": [
      {
        "object": "block",
        "type": "paragraph",
        "paragraph": { "rich_text": [{ "type": "text", "text": { "content": "{{ $json.text }}" } }] }
      }
    ]
  }
}

Step 6: Save, Activate, and Monitor

1. Click Save (top-right) — name it "Daily AI Research Digest"
2. Toggle the Active switch (top-left) to ON
3. The workflow is now running on schedule. n8n shows execution history under the Executions tab

Manual test: Click Test Workflow to run the full pipeline immediately without waiting for the schedule.

Common Pitfalls and Fixes

Pitfall 1: "401 Unauthorized" from DeepSeek

Cause: Wrong Base URL or missing /v1 suffix. Fix: In the OpenAI Chat Model credentials, set Base URL to https://api.deepseek.com/v1 — the trailing /v1 is required even though the OpenAI node appends /chat/completions automatically.

Pitfall 2: AI node returns empty/truncated output

Cause: The {{ $json.articles }} expression passes the raw array object — DeepSeek may parse it incorrectly. Fix: Insert an Item Lists → Summarize node before the AI Chain, or use a Code node to pre-format the articles into a string:

const articles = $input.first().json.articles;
const formatted = articles.map((a, i) =>
  `${i+1}. ${a.title}\n   ${a.description}`
).join('\n\n');
return [{ json: { formatted } }];

Pitfall 3: Slack message exceeds character limit

Cause: DeepSeek generates a long digest, Slack limits to 4000 chars. Fix: In the Basic LLM Chain prompt, add: "Keep the total output under 3500 characters." Or split long messages using a Code node.

Pitfall 4: Schedule trigger misses the window

Cause: n8n's schedule is based on the server's UTC time, not your local time. Fix: Set the trigger hour in UTC. If you're in EST (UTC-5) and want 8 AM, set hour to 13 (8+5).

Pitfall 5: Docker container stops after reboot

Cause: docker run doesn't auto-restart containers. Fix: Re-run with --restart unless-stopped:

docker run -d --restart unless-stopped --name n8n -p 5678:5678 ...

Performance and Cost

With DeepSeek's pricing, a daily digest of 10 articles costs roughly:

Yes, three cents per month for a fully automated AI research assistant.

Next Steps

Now that you have a working pipeline, extend it:

• Add filtering: Use an IF node to skip articles below a relevance threshold
• Multi-source: Chain multiple HTTP Request nodes for RSS + NewsAPI + Reddit
• Vector search: Store digests in Pinecone or Qdrant for semantic search across past issues
• Custom alerts: Add a Webhook trigger that fires when a topic keyword appears (e.g., "Claude Code" or "Hermes Agent")

This tutorial was built with n8n self-hosted, DeepSeek API, and the n8n OpenAI Chat Model node. The same pattern works with Claude API, OpenAI GPT-4o, Gemini, or any OpenAI-compatible endpoint — just swap the Base URL and model name in the credentials.