[New Node] DuckDuckGo Search Integration for n8n

[Sam_Nodehi]

Hi everyone!

I’m excited to share a new community node I developed: DuckDuckGo Search for n8n

GitHub: https://github.com/samnodehi/n8n-nodes-duckduckgo
npm: https://www.npmjs.com/package/n8n-nodes-duckduckgo-search

Features:

• Perform fast DuckDuckGo web searches directly from your n8n workflows
• Simple configuration, no API key required
• Supports search result parsing

You can install it using:

npm install n8n-nodes-duckduckgo-search

[Sanjay_B]

I can’t able to see this node in n8n after installation

can I able to add this node in AI agent node

[Sam_Nodehi]

Thanks for your question

This node currently works only as a standalone node in n8n and cannot be directly connected to the AI Agent as a tool.
To use it, simply search for duckduckgo in the “Add Node” panel, and you’ll find it listed there.

[Sanjay_B]

Actually I installed and added it still when I try to make 70 queries parallelly, 50% of them not getting results

how to fix it (any methods to get proper results always]

and also if It’s possible to add in agent node means that also be useful

[Sam_Nodehi]

The reason you’re getting partial results when running 70 parallel queries is due to DuckDuckGo’s internal anti-bot and caching system, it starts serving empty cached responses when requests come too frequently or in large batches.

I recommend adding a short randomized delay (for example 3–8 seconds) between consecutive requests or batching them in smaller groups (e.g., 10–15 queries per batch). The next version of this node will include an adaptive backoff mechanism to handle this automatically when consecutive empty responses are detected.

Regarding AI Agent integration: I’m preparing an update that will make this node available directly in the AI Agent’s “Tools” dialog. I’ll share detailed instructions and a version tag once it’s published.

[Sam_Nodehi]

HUGE UPDATE v31.0.1: AI Agent Ready & Production-Grade Reliability!

---

Hi everyone!

I’m thrilled to announce a major, foundational update for the DuckDuckGo Search node, version v31.0.1, which is now available.

This release completely re-engineers the node for production-grade stability and full integration with n8n’s AI Agent.

Key New Features

• AI Agent Ready: The node is now fully selectable as an AI Agent Tool! You can pick it directly from the “Tools” dialog in the AI Agent node for advanced AI workflows.

• Production-Grade Reliability: A new robust reliability system has been implemented to gracefully handle rate limits and prevent empty results under heavy/parallel loads. This includes an adaptive backoff, jitter, and a circuit breaker.

• Advanced Configuration: Power users can now fine-tune all reliability settings (backoff thresholds, concurrency limits, etc.) to customize performance for specific, high-volume workloads.

This update transforms the node from a simple tool into a reliable component for serious AI tasks and high-volume automations.

---

A Quick Note on this Release

This version, v31.0.1, is the official, stable release. A previous version (v31.0.0) was briefly available but had a critical bug where these new reliability features were not activating. This has been 100% fixed and verified (with 141 automated tests) in v31.0.1.

Installation

Please ensure you install the latest, stable version:

Bash

npm install n8n-nodes-duckduckgo-search@latest

I’m excited for you to try the new capabilities and, as always, all feedback is welcome!

[Sam_Nodehi]

Update: v32.5.2 is now available

I’ve shipped a cleanup and quality update for the DuckDuckGo Search community node for n8n.

This update focuses on making the node more predictable, cleaner for AI Agent workflows, and more honest about what the runtime actually does.

What changed recently:

• Cleaner node UI: removed unused / misleading options that did not affect runtime behavior

• No API key required for DuckDuckGo search

• No outbound telemetry

• Works as an n8n AI Agent tool

• Web, Image, News, and Video search supported

• Web Search output is cleaner: sponsored/ad redirect URLs are filtered

• DuckDuckGo redirect URLs are normalized to final destination URLs

• Image Search now includes a useful source page URL

• News fallback behavior is safer and more relevant

• News/Video fallback results are clearly marked with isFallback: true

• Better error messages for image VQD issues, 403 blocks, and parser failures

Latest version:

npm install n8n-nodes-duckduckgo-search@latest

Links:

GitHub:
https://github.com/samnodehi/n8n-nodes-duckduckgo

npm:
https://www.npmjs.com/package/n8n-nodes-duckduckgo-search

I also rewrote the README to better match the current runtime behavior and output fields.

If you use this node inside AI Agent workflows, I’d recommend updating to the latest version.

[Kaushik_Ghosh1]

Can anyone help why I get 0 output results for this simple query

[Sam_Nodehi]

Hi, thanks for reporting this.

I tested the same query on the latest package version (32.5.2) and it returns results correctly on my side. For example, the query:
Latest FDA updates on Class II medical devices on 2025

returns 10 web results, including Federal Register / FDA-related pages.

A couple of things to check:

1. Please make sure you are using the latest version of the community node:

npm install n8n-nodes-duckduckgo-search@latest

2. Restart n8n after updating the community package.
3. In n8n, open the node details and confirm the package version shows 32.5.2 or newer.
4. Try running the query directly as plain text first, without expressions in Search Query or Maximum Results, just to rule out an expression/input issue.
5. If you are using Docker, make sure the updated package is installed inside the actual n8n container/volume that n8n is running from.

If it still returns 0 after updating and restarting, please share:

• your installed package version
• your n8n version
• whether you are using Docker or desktop/cloud
• the exact node parameters from the JSON view

Then I can check whether it is an environment/version issue or a real parser issue.

[nguyenthieutoan]

For those who want to use this as an AI Agent tool - since the node can’t be dropped directly into the agent’s tool connector yet, a clean workaround is to wrap it in a sub-workflow:

1. Create a separate workflow with a “When called by another workflow” trigger that accepts a query parameter
2. Drop the DuckDuckGo node there and pass {{ $json.query }} into the search query field
3. Return the results
4. In your main AI agent workflow, add an “Execute Workflow” node as an agent tool and point it to that sub-workflow

The agent then treats the sub-workflow as a callable tool and passes search queries through it. Works reliably and lets you keep the DuckDuckGo node in a standalone position while still plugging into the agent flow.

[Sam_Nodehi]

Thanks for sharing the workaround.

Just a quick update: in the latest versions of the DuckDuckGo node, it is already AI Agent tool-compatible and can be used directly as an n8n AI Agent tool.

So for most users, the sub-workflow workaround is no longer required.

If the node does not appear in the AI Agent tool connector, please check:

1. You are using the latest package version:

npm install n8n-nodes-duckduckgo-search@latest

2. Restart n8n after updating the community node.

3. For some self-hosted/community-node setups, make sure community nodes are allowed as AI tools:

N8N_COMMUNITY_PACKAGES_ALLOW_TOOL_USAGE=true

Your sub-workflow approach is still a valid fallback/workaround, especially for older n8n versions or restricted setups, but the node itself is now designed to work directly as an AI Agent tool.

[Sam_Nodehi]

Big update — Page Content, Extract-from-URL, Instant Answers & metadata

Hi everyone A lot has landed since launch. Here’s everything new — all free, no API key, no paid dependencies (requests go only to DuckDuckGo, except the two opt-in page-fetch features below).

Fetch Page Content — Web & News Search (opt-in, off by default)
Goes beyond DuckDuckGo’s snippet and pulls the main readable text of each result page into pageContent. Three-tier extraction (Mozilla Readability → DOM heuristic that strips menus → regex fallback); one bad page never fails the search.

• Number of Results to Fetch — how many top results to enrich
• Max Content Length — cap per page (sets pageContentTruncated)
• Page Fetch Timeout — per-page timeout; errors land in pageContentError
• Include Page Metadata — adds pageTitle, pageAuthor, pagePublished, pageExcerpt, pageSiteName

Extract Page Content — new operation
Give it any URL → clean main text (content) + optional metadata. A “read any page” tool for when your agent already has a link.

{ "operation": "extractContent", "url": "https://example.com/article",
  "extractOptions": { "includePageMetadata": true, "pageContentMaxLength": 3000 } }

Instant Answer — new operation
Direct answers, Wikipedia-style abstracts, definitions & related topics via DuckDuckGo’s official free, no-key Instant Answer API — a quick factual answer, no scraping.

{ "operation": "instantAnswer", "instantAnswerQuery": "DuckDuckGo" }

→ returns heading, abstract, abstractSource/abstractURL, answer/answerType, definition, relatedTopics[], type

Privacy: the two page-fetch features are the only ones that contact third-party sites, so they’re opt-in. With them off, requests go only to DuckDuckGo — unchanged behaviour.

Under the hood (32.6 → today): big internal cleanup (dead code removed, slimmer deps, proper MIT license, fixed package entry), unified locale/region (wt-wt default), more tests + CI hygiene, and dependencies kept current.

Update: Community Nodes will offer it, or:

npm i n8n-nodes-duckduckgo-search@latest

Full notes: CHANGELOG → n8n-nodes-duckduckgo/CHANGELOG.md at main · samnodehi/n8n-nodes-duckduckgo · GitHub · Releases → Releases · samnodehi/n8n-nodes-duckduckgo · GitHub

Feedback & bug reports always welcome — thanks for the support!