n8n-nodes-deep-research

n8n-nodes-deep-research

Free, zero-API-key research node for n8n — search the web, Reddit, Hacker News, Wikipedia, news feeds, scrape any page, and run deep multi-source research, all at zero cost.

Features

Operation	Source	Auth Required	Description
Web Search	DuckDuckGo	No	General web search
Reddit Search	Reddit JSON API	No	Search posts with sort & time filters
Hacker News	HN Algolia API	No	Search stories, comments, Ask/Show HN
Wikipedia	Wikipedia API	No	Search articles, get extracts, any language
News / RSS	Google News, BBC, TechCrunch, Custom	No	Search news via RSS feeds
GitHub	GitHub Public API	No	Search repos, issues, code, users
Scrape Web Page	Any URL	No	Fetch page, extract text, links, metadata
Deep Research	ALL sources combined	No	Aggregates from every source in parallel

Install

Option 1: Install via npm (recommended)

# In your n8n custom nodes directory
cd ~/.n8n/custom
npm install n8n-nodes-deep-research

Then restart n8n.

Option 2: Install globally and link

# Clone or download this repo
git clone https://github.com/your-username/n8n-nodes-deep-research.git
cd n8n-nodes-deep-research
npm install
npm run build

# Link globally
npm link

# In your n8n custom nodes directory
cd ~/.n8n/custom
npm link n8n-nodes-deep-research

Option 3: Docker

Add to your docker-compose.yml:

volumes:
  - ./n8n-nodes-deep-research:/usr/local/lib/node_modules/n8n-nodes-deep-research

Or in your Dockerfile:

COPY n8n-nodes-deep-research /usr/local/lib/node_modules/n8n-nodes-deep-research
RUN cd /usr/local/lib/node_modules/n8n-nodes-deep-research && npm install && npm run build

Option 4: n8n Community Install (if published to npm)

In n8n, go to Settings → Community Nodes → Install and search for deep-research.

Usage

1. Open any n8n workflow
2. Add a Deep Research node
3. Select your operation from the dropdown
4. Enter your search query
5. Configure any source-specific options (sort, time filter, etc.)
6. Run the workflow

Example: Search Reddit for product feedback

Operation: Reddit Search
Query: "best mechanical keyboard 2025"
Max Results: 15
Sort By: Top
Time Filter: Past Month

Example: Deep Research on a topic

Operation: Deep Research
Query: "quantum computing applications healthcare"
Max Results: 25

This will run searches across DuckDuckGo, Reddit, Hacker News, Wikipedia, and Google News simultaneously and return all results.

Example: Scrape a web page

Operation: Scrape Web Page
URL: https://example.com/article
Extract Links: true
Max Text Length: 20000

Output Format

Each result is returned as a separate item with these common fields:

{
  "title": "Result Title",
  "url": "https://example.com/page",
  "snippet": "Brief description or excerpt...",
  "source": "duckduckgo|reddit|hackernews|wikipedia|news-google|github"
}

Source-specific extra fields:

Reddit: redditScore, redditAuthor, redditSubreddit, redditNumComments, redditCreatedUtc

Hacker News: hnPoints, hnAuthor, hnNumComments, hnCreated, hnObjectId

Wikipedia: wikiWordcount, wikiTimestamp

GitHub: githubStars, githubForks, githubLanguage, githubTopics

News/RSS: publishedAt, author

Scrape: description, ogTitle, ogDescription, ogImage, textContent, textLength, links, linksCount

Return as Single Item

Enable Additional Fields → Return as Single Item to get all results wrapped in a single item:

{
  "operation": "deepResearch",
  "totalResults": 25,
  "results": [
    { "title": "...", "url": "...", "source": "..." },
    ...
  ]
}

Development

# Install dependencies
npm install

# Start development with hot reload
npm run dev

# Build for production
npm run build

# Lint code
npm run lint

# Lint and auto-fix
npm run lint:fix

How It Works (No API Keys Needed)

• DuckDuckGo: Parses their HTML search page (no API key needed)
• Reddit: Uses the public reddit.com/search.json endpoint
• Hacker News: Uses the Algolia HN API (hn.algolia.com)
• Wikipedia: Uses the MediaWiki Action API
• News: Parses RSS/Atom feeds from news sites
• GitHub: Uses the unauthenticated public Search API (60 requests/hour)
• Scraping: Fetches raw HTML and extracts content

Limitations

• DuckDuckGo HTML parsing may break if DDG changes their page structure
• Reddit's public API has rate limits (without OAuth)
• GitHub unauthenticated API: 60 requests/hour
• Web scraping depends on the target page's structure
• No authentication means some rate-limited APIs may be slower

License

MIT