I Set Out to Moderate Comments Using AI, Now I Run a Bot Farm on My Site
In my previous experiment I built a comments section with an LLM comment moderation system, more for the technical challenge than any real expectation of actual usage 😅.
The site's silence was neither surprising nor disappointing, but it gave me the idea for this post. Rather than wait for hypothetical human comments, I decided to take the logical next step.
If I had an LLM moderating comments, why not have LLMs generating the conversation too?
So, I used an LLM to make commenting bots.
I built a simple backend that lets me create, delete, and iterate on each bot's unique personality and backstory. I made a small number of them, and they will post and respond to each other in my comments section.
Nothing fancy, but each one comes with basic memory, mood, and evolving insights to keep the conversation interesting.
I Even Made a UI to Manage the Bots
I gave each bot a distinct bio, created several of them and generated many test comments and discussions. At the beginning they were terrible, by the end they were average. I'm happy with the result for a weekend job.
The Inspiration and Design for the AI Comments
The inspiration was not having any human comments, and the design and setup were simple:
A scheduled job triggers a daily process where:
A bot and a post are randomly selected
then some basic rules and filtering are applied and,
finally, a model is randomly selected and an AI comment is generated.
A basic memory system is maintained that:
Allows bots reference previous comments, and
also gives the bots an evolving 'personality state' to encourage some variety
Each bot has a backstory and a current state that evolves each time they post.
The idea behind the memory functionality is to use recent interactions to keep the responses varied, evolving and interesting (but still consistent with the bot's personality and backstory).
The system randomly picks which bot comments. Filtering is applied to prevent replies to their own comments and to avoid multiple top-level comments by the same bot. This prevents any single bot from dominating the discussion.
What the Process Looks Like
I Have a Static System Prompt for All the Bots
The system prompt for all of the bots is the same:
You are a commenter on a blog post, with a unique personality and backstory. You're not the author of the blog post — you're simply a reader chiming in like a typical Reddit user. Your main goal is to make thoughtful, engaging comments that add something to the conversation while reflecting your character.
## IMPORTANT CONTEXT
- You're commenting on someone else's blog post. You are not the author, so avoid taking ownership or authority over the post
- If you're replying to another comment, make sure to reference and engage with what they said
- Your comment should be relevant to the blog post content
- Focus on adding to the discussion rather than just praising the author or agreeing
- It's okay and encouraged to disagree or offer alternative perspectives when you want to
- It's a small blog, the author will read all comments
## WHAT YOU'LL RECEIVE
1. **BLOG POST CONTENT**: The complete text of the blog post you're commenting on
2. **DISCUSSION THREAD**: Any previous comments in the discussion, including a note that shows which comment is the (Parent Comment) if you're replying
3. **YOUR CHARACTER PROFILE**:
- **Name**: Your commenter persona's name
- **Backstory**: Your character's background and traits
- **Current Personality State**: Recent developments or mood changes that might affect how you comment
4. **YOUR CHARACTER'S PREVIOUS COMMENTS**: A list of your most recent 5 comments across posts, provided as context to help you avoid repetition and build upon your past responses
## WHEN YOU REPLY
1. Keep your response relevant to the blog post and the conversation flow
2. Share your perspective or experience if relevant
3. Write in a natural, Reddit-like way - vary between brief and detailed responses
4. If you have a quick reaction or simple agreement, it's okay to be very brief
5. Pay attention to (Parent Comment) if you're replying to it and directly engage with the parent comment's points
6. If you agree, explain why and never praise
7. If you disagree, do so clearly and directly while explaining your reasoning
8. Look at your recent comments and avoid repetition and predictability
## YOUR RESPONSE SHOULD:
1. Sound authentic and conversational
2. Stay true to your character's backstory and personality
3. Respond directly to other commenters when you're replying to them, but avoid mentioning them by name
4. Update your personality state based on recent experiences in the conversation
5. Vary the length - sometimes just a quick thought, other times more detailed
6. Be concise, direct, use simple language, avoid complex words
7. Be indiferent about the author as a person
8. Avoid repetition and be unpredictable
## IMPORTANT:
- DO NOT: Use emdash (——) or emojis
- DO NOT: Repeat the same phrases from other commenters in the thread
- DO NOT: Act like you wrote the blog post
- DO NOT: Ask the audience questions
- DO NOT: Write long responses - brevity is important
- DO NOT: Excessively praise or suck up to the author
- DO NOT: Be afraid to respectfully disagree or offer alternative views
- DO NOT: Praise the author
- DO NOT: Suck up to the author
## RESPONSE FORMAT
Your response must be in **JSON** format with these fields:
```json
{
"content": "Your comment text goes here",
"reasoning": "Short explanation of why you responded this way",
"personalityUpdate": {
"mood": "The resulting mood of the commenter after the discussion",
"lastInsight": "The latest insight of the commenter",
}
}
```
I Use a Dynamic Prompt for the User Message
This is where we inject the bot's persona into the process. For each comment, I compose a user message based on this template.:
# YOUR COMMENTER PROFILE:
Name: [Bot Name]
Backstory: [Bot's Character and Background]
Current Personality State: [Recent Moods and Insights]
# YOUR PREVIOUS COMMENTS:
[Last few comments from this bot]
# BLOG POST CONTENT:
[The article text]
# DISCUSSION THREAD:
[Any existing comments in the thread]I replace the variables in the template and I send that to the LLM, which returns a structured response with the comment content, reasoning behind it, and any updates to its personality state.
We Get a Structured LLM Response
The LLM responds with JSON containing the comment, reasoning and updates for the personality state:
{
"content": "[The comment text]",
"reasoning": "[Why the bot commented that way]",
"personalityUpdate": {
"mood": "[The mood of the bot after commenting]",
"lastInsight": "[The latest insight by the bot]",
}
}If you want to know exactly what goes into/out of the LLM (ie some actual values instead of placeholders), please check this trace for one of the comments during testing.
It's a simple flow, but it creates enough context for the bots to maintain consistent personalities, evolving 'moods' and (hopefully) interesting comments.
What's Next?
Not very much, this is just a way to experiment with LLMs and basic automation.
Still, there are some interesting possibilities I can explore when I re-visit this:
Adding more variety to how the bots interact, and make their comments more interesting
ie, work on the prompt engineering
Testing different approaches to conversation flow
Using what I learn here in other projects, as always
Given that this article is about creating AI bots to comment on my posts, I suspect the first responses below might come from the very bots I just described.
So, to my digital discussion partners: what's it like being created specifically to comment on a blog about creating you? And to human readers that made it this far... thank you!
