How I Organize AI Chat Sessions Automatically | Hermes Session Cleaner Tutorial

Never Lose Track of Your AI Conversations Again

The Hermes Agent Session Cleaner is the answer to a problem every AI power user runs into. If you use Hermes Agent, or any AI tool that generates conversation logs, you know how fast they pile up. Raw JSON files. .jsonl streams from Telegram. CLI checkpoint files. Dozens of sessions a day, each one a file you have to dig through if you want to find something you talked about yesterday.

I built a solution for this. It is called the Session Cleaner, and it runs automatically in the background three times a day. If you are new to autonomous AI in general, my post on why AI agents like Hermes are the future is a good primer before you dig in.

Watch the Walkthrough

What the Hermes Agent Session Cleaner Does

The Session Cleaner is a cron-powered script that automatically extracts clean, readable markdown from hundreds of Hermes session files and organizes them by date and platform on your Desktop.

  • Scans Hermes session files across every platform you run, including CLI checkpoints and Telegram streams
  • Strips out tool calls, reasoning traces, and system prompts so the output reads like a normal chat transcript
  • Merges fragmented checkpoint files into a single continuous conversation per session
  • Names every file with the date, platform, and time slot so you can find any conversation in seconds
  • Runs three times a day on cron with zero manual triggering, so the archive stays current automatically

The output lands in ~/Desktop/Sessions/Main Chat/YYYY-MM-DD/ with filenames like CLI-2026-05-12-Morning.md and Tele-2026-05-12-Afternoon.md. Open the folder, and every conversation you had today is right there in plain markdown.

Why the Hermes Agent Session Cleaner Matters

Hermes Agent is a persistent, autonomous AI. It works across sessions, learns from experience, and carries context forward, similar to the trend I covered in my May 2026 AI Marketing Dispatch on Hermes Agent. But that context is only useful if you can actually look back at what happened.

Without the Session Cleaner, reviewing past conversations means opening raw JSON files full of tool call metadata, reasoning traces, and system prompts. With it, you get clean markdown that reads like a chat log, exactly what you need for reference, sharing with clients, or writing follow-up content.

It is particularly useful if you:

  • Reference past AI conversations to pick up work where you left off the day before
  • Share session logs with clients or teammates without exposing raw JSON internals
  • Pull material from chat sessions into blog posts, emails, documentation, or training content
  • Run Hermes across multiple devices and want one searchable, date-organized archive in one place

How the Time Slots Work

The day is split into three windows, each with its own cron job and output file:

SlotHoursRuns AtFile Suffix
MorningMidnight to 12 PMNoon-Morning.md
Afternoon12 PM to 5 PM5 PM-Afternoon.md
Evening5 PM to MidnightMidnight-Evening.md

Each run regenerates the file from scratch, so continuations from later in the day blend seamlessly into the same section. No duplicate entries, no stale data.

Get Started

The Hermes Agent Session Cleaner is built into Hermes Agent as a skill. Load it with skill_view('session-cleaner') inside your Hermes CLI, or check the cron jobs with cronjob list to see the three automated runs. You can read more about Hermes itself on the Nous Research site.

If you are not using Hermes Agent yet, install it with one command:

curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash

Once installed, run hermes setup, connect your provider, and you have a persistent AI agent that organizes its own conversations.


Who this is for: Hermes Agent users, AI power users, agency owners running AI workflows, marketers automating with AI, and anyone who has stared at a folder full of JSON files wondering where that conversation went.

Leave a Comment

Scroll to Top