# n8n research agent: connect the tools and check the evidence

Trace the user question through the mapped fields and returned sources before judging the final essay.

[HTML](https://lukejbyrne.com/library/guides/6ca4aoqboao/)
Content language: en
Author: Luke Byrne
Video published: 2025-06-18
Guide published: 2026-11-02

[Watch the video](https://www.youtube.com/watch?v=6ca4aoqbOao)
Video language: English

## Check the node and model setup

The June 2025 lesson builds a research agent with Perplexity and Google Gemini inside n8n. Its goal is a substantial output such as an essay with references. The title’s ten-times-smarter claim is promotional framing, not a measured intelligence result reported by this companion.

Perplexity setup begins at 1:08, version compatibility at 2:18, Gemini API setup at 3:43 and the Perplexity key at 5:36. Keep the installed n8n version and the selected model route visible when following this historical setup; current node names and access terms may differ.

1. Follow the version and connection sections before running a research task.
2. Configure the intended Gemini and Perplexity access.
3. Inspect a small tool response to distinguish a connection failure from an instruction problem.

## Map the question instead of relying on a plausible default

Agent instructions are covered at 7:56 and direct chat-input mapping at 10:01. Check which field actually reaches the agent. A fluent report can answer the wrong question if the workflow passes a placeholder, previous input or unintended value.

Describe the desired research scope and the source evidence needed to support it. The source promises prompts and JSON templates elsewhere but does not include their contents here, so no exact recipe is reconstructed from the title.

1. Use one short research question with a clear expected scope.
2. Inspect the mapped input at the agent step.
3. Compare the returned report with that exact question before expanding it.

## Read references as part of the result

The live test begins at 13:45. Open a sample of the cited sources and compare the report’s claims with their content. A bibliography is useful when the references genuinely support the statements, not simply because links appear below a long answer.

Keep uncertainty and missing evidence visible, and target a specific weakness when revising the instructions. The description mentions API efficiency and troubleshooting but supplies no measured cost comparison or proof that the agent is more accurate than every alternative. Evaluate the research through the question, retrieved material and supported conclusion.

## The takeaway

Correct input mapping and relevant references matter more than report length or the title’s intelligence claim.

## Workflow resources

- [ailuke.short.gy](https://ailuke.short.gy/6ca4aoqbOao1)
- [ailuke.short.gy](https://ailuke.short.gy/6ca4aoqbOao2)
- [datacamp.pxf.io](https://datacamp.pxf.io/yqr3jb)
- [newsletter.lukejbyrne.com](https://newsletter.lukejbyrne.com)
- [linkedin.com](https://linkedin.com/in/lukejbyrne)
- [x.com](https://x.com/thelukejbyrne)
- [datacamp.pxf.io](https://datacamp.pxf.io/gO6WRB)
- [datacamp.pxf.io](https://datacamp.pxf.io/XmnX4a)
- [datacamp.pxf.io](https://datacamp.pxf.io/Bny25L)
- [datacamp.pxf.io](https://datacamp.pxf.io/aOAWNZ)
- [datacamp.pxf.io](https://datacamp.pxf.io/AP3Eg1)
- [datacamp.pxf.io](https://datacamp.pxf.io/GKVZbk)
- [datacamp.pxf.io](https://datacamp.pxf.io/Z6KyV1)

## Chapters

- [0:00 – Perplexity AI Agent: Game Changer](https://www.youtube.com/watch?v=6ca4aoqbOao&t=0s)
- [1:08 – Setting Up Perplexity in n8n](https://www.youtube.com/watch?v=6ca4aoqbOao&t=68s)
- [2:18 – Ensure n8n Version Compatibility](https://www.youtube.com/watch?v=6ca4aoqbOao&t=138s)
- [3:43 – Google Gemini API Setup](https://www.youtube.com/watch?v=6ca4aoqbOao&t=223s)
- [5:36 – Connecting Perplexity with API Key](https://www.youtube.com/watch?v=6ca4aoqbOao&t=336s)
- [7:56 – Defining Agent Instructions for Accuracy](https://www.youtube.com/watch?v=6ca4aoqbOao&t=476s)
- [10:01 – Directly Mapping Chat Inputs](https://www.youtube.com/watch?v=6ca4aoqbOao&t=601s)
- [13:45 – Live Agent Test with Perplexity](https://www.youtube.com/watch?v=6ca4aoqbOao&t=825s)

## Sources and limitations

Written companion prepared 2 November 2026 from this exact published video description and its chapter sequence. The full recording was not independently watched or transcribed for this article, and the products and reported results were not retested. Checklists are editorial adaptations of the documented workflow. Model availability, interfaces, prices and offers may have changed since publication; use the retained original resources for current requirements.

[Original video](https://www.youtube.com/watch?v=6ca4aoqbOao)
