Im trying to use some irc logs to create a workflow to analyse mood of conversations using AI agents, por example, if on day 01 of january the mod was friendly or not, and so on.
But i searched and didnt found any way to do this (please keep in mind im kind of new, so i could miss something).
The site with the logs is this https://irclogs.ubuntu.com/, they have some kind of sorting by year, then month, then day, like a folder for 2004, then 01 for january, then 01 for day 1.
You guys have any idea how to do this? Like, how to extract from that html or txt files then group them them use ai to analyse it and save? Is it even possible for n8n?
Yes, it’s possible in n8n. Here’s how to do it step by step: Step 1: Read the logs from irclogs.ubuntu.com
The logs are organized as follows: https://irclogs.ubuntu.com/2004/01/01/
Each day has a .txt or .html file, for example: https://irclogs.ubuntu.com/2023/12/25/%23ubuntu.txt
How to extract it in n8n?
Use an HTTP Request node:
Method: GET
Dynamic URL based on date.
Header: None special, this is public content.
Optional: Use a Function node to build URLs by date range.
Step 2: Clean and group the content
If it’s HTML:
Use HTML Extract to extract the chat content.
If it’s .txt:
Use a Code node to parse lines by user and message.
Step 3: Analyze sentiment (mood)
Given that you have access to Gemini 2.0, you could:
Use an AI agent node in n8n (such as the AI Agent node with HTTP Request to the Gemini endpoint, if you have an API).
Or simply use OpenAI, Grok, or Claude if they’re already integrated.
Step 4: Save the results
You can use:
Google Sheets (Google Sheets node)
Airtable
Notion
I visualize the flow like this:
Start
HTTP Request → Get log by date
Code → Parse messages
AI Agent → Mood analysis
Set → Format output
Database/Sheet/File → Save result