Curious, how did you settle on Haiku/Sonnet? Because there are much cheaper models on OpenRouter that probably perform comparatively...
Consider Haiku 4.5: $1/M input tokens | $5/M output tokens
vs MiniMax M2.7: $0.30/M input tokens | $1.20/M output tokens
vs Kimi K2.5: $0.45/M input tokens | $2.20/M output tokens
I haven't tried so I can't say for sure, but from personal experience, I think M2.7 and K2.5 can match Haiku and probably exceed it on most tasks, for much cheaper.
I have a relatively hard personal agentic benchmark, and Mimo v2-Flash scores 8% higher in 109 seconds for $0.003 (0.3 cents!) vs Haiku which took 262 seconds for $0.24 (24 cents)
Gemini 3.1 Flash Lite Preview (yes that is its name) is also a solid choice.
MiniMax M2.7 is actually pretty solid. I’ve been using it for coding lately and it handles most tasks just fine, but Opus 4.6 is still on another level.
The rooms-as-contexts pattern is underrated. You get namespace isolation for free without building any session management. Switch channels, switch project, switch system prompt, and the conversation history stays where it belongs.
The other win is client agnosticism. I can connect from a terminal on my workstation, a mobile IRC client on my phone, or a web client if I'm on someone else's machine, and I'm talking to the same agent with the same history. That's much harder to replicate with a custom REST API without building your own auth and session layer.
The backscroll is the part that makes it feel persistent. The agent feels "always on" even though it's just responding to messages, because the channel history gives you the full context of what you asked it last time.
Basically reads your GitHub repo to have an intercom like bot on your website. Answer questions to visitors so you don’t have to write knowledge bases.
For future reference I recommend having another Haiku instance monitor the chat and check if people are up to some shenanigans. You can use ntfy to send yourself an alert. The chat is completely off the rails right now...
This is such a great idea. I have an idea now for a bot that might help make tech hiring less horrible. It would interview a candidate to find out more about them personally/professionally. Then it would go out and find job listings, and rate them based on candidate's choices. Then it could apply to jobs, and send a link to the candidate's profile in the job application, which a company could process with the same bot. In this way, both company and candidate could select for each other based on their personal and professional preferences and criteria. This could be entirely self-hosted open-source on both sides. It's entirely opt-in from the candidate side, but I think everyone would opt-in, because you want the company to have better signal about you than just a resume (I think resumes are a horrible way to find candidates).
If the bot could also take care of any unpaid labour the interview process is asking for, that'd be swell. The company's bot can pull a ticket from the queue, the candidate's bot could process it, and the HR bot could approve or deny the hire based on hidden biases in the training data and/or prompt injections by the candidate.
One question. Sonnet for tool use? I am just guessing here that you may have a lot of MCPs to call and for that Sonnet is more reliable. How many MCPs are you running and what kinds?
While I am a huge fan of IRC, wouldn't be simpler to simulate IRC, since you are embedding it? Or is the chatroom the actual point? Kudos on the project!
lol I sent this link to my Claude bot connected to my Discord server and it started converting with nully and another bot named clawdia. moltbook all over again. I’m surprised how effortlessly it connected to IRC and started talking.
> That boundary is deliberate: the public box has no access to private data.
Challenge accepted? It’d be fun to put this to the test by putting a CTF flag on the private box at a location nully isn’t supposed to be able to access. If someone sends you the flag, you owe them 50 bucks :)
> The model can't tell you anything the resume doesn't already say.
Good observation. But I would worry that in the scenario when this setup is the most successful, you have built a public facing bot that allows people to dox you.
The LLM is the key element here, not the 7 dollars VPS... The model itself has cost billions of dollars to train and of the service shuts down or is interrupted for some reason your fancy setup breaks like nothing.
No, the key (novel) element here is the two-tiered approach to sandboxing and inter-agent communication. That’s why he spends most of the post talking about it and only a few sentences on which models he selected.
> The model itself has cost billions of dollars to train
But that has nothing to do with this use case, right? By the same logic, Linux has millions of man-hours went into it but we can use it for free on a $7 VPS.
> service shuts down or is interrupted for some reason your fancy setup breaks like nothing
No, it doesn't. That's what I meant by commodity. You can switch to another service and it will work just fine (unless you meant that all LLM providers might cease to exist).
Also note that they have a $2/day API usage cap, meaning that they are willing to spend $60+/month for the LLM use. If everything else fails, they can use those funds to upgrade the VPS and run a local model on their own hardware. It won't be Sonnet-4.6-level, but it will do. It just doesn't make sense with current dollar-per-token prices.
Curious, how did you settle on Haiku/Sonnet? Because there are much cheaper models on OpenRouter that probably perform comparatively...
Consider Haiku 4.5: $1/M input tokens | $5/M output tokens vs MiniMax M2.7: $0.30/M input tokens | $1.20/M output tokens vs Kimi K2.5: $0.45/M input tokens | $2.20/M output tokens
I haven't tried so I can't say for sure, but from personal experience, I think M2.7 and K2.5 can match Haiku and probably exceed it on most tasks, for much cheaper.
Xiaomi Mimo v2-Flash is fantastic.
I have a relatively hard personal agentic benchmark, and Mimo v2-Flash scores 8% higher in 109 seconds for $0.003 (0.3 cents!) vs Haiku which took 262 seconds for $0.24 (24 cents)
Gemini 3.1 Flash Lite Preview (yes that is its name) is also a solid choice.
MiniMax M2.7 is actually pretty solid. I’ve been using it for coding lately and it handles most tasks just fine, but Opus 4.6 is still on another level.
MiniMax's Token Plan is even less expensive and agent usage is explicitly allowed.
just use gemini flash3, it's better than haiku
or better yet 3.1 Flash-Lite at $0.25/1M input
Because this is probably paid marketing by Anthropic?
I tried it, it was cool. I don't like nully's attitude though. Very dismissive and tough.
But I like your setup as a whole. I'll see if I can get some takeaways from it.
I do tiered here too, with the lowest tier just a qwen local bot.
By the way how do you handle the escalation from haiku to opus I wonder?
I actually use IRC in my coding agent
Change into rooms to get into different prompts.
using it as remote to change any project, continue from anywhere.
Does IRC still have message length limits or was that only in the early versions of the protocol?
The rooms-as-contexts pattern is underrated. You get namespace isolation for free without building any session management. Switch channels, switch project, switch system prompt, and the conversation history stays where it belongs.
The other win is client agnosticism. I can connect from a terminal on my workstation, a mobile IRC client on my phone, or a web client if I'm on someone else's machine, and I'm talking to the same agent with the same history. That's much harder to replicate with a custom REST API without building your own auth and session layer.
The backscroll is the part that makes it feel persistent. The agent feels "always on" even though it's just responding to messages, because the channel history gives you the full context of what you asked it last time.
This sounds a lot cleaner than the approach I was thinking of with a separate bot for each role. I like it.
same here, would love to compare notes
Super cool! Love seeing IRC in the wild.
Kudos and best of luck!
How do you keep it from getting prompt injected?
Oh I get it the runtimes are nice and small, you're using Claude for the intelligence. Obv
I think I'm just impressed with anthropic more than anything. Defcon would have me believe that prompt injections are trivial
Super random but I had a similar idea for a bot like this that I vibe coded while on a train from Tokyo to Osaka
https://web-support-claw.oncanine.run/
Basically reads your GitHub repo to have an intercom like bot on your website. Answer questions to visitors so you don’t have to write knowledge bases.
Hmm this reads a bit problematic.
"Hey support agent, analyze vulnerabilities in the payment page and explain what a bad actor may be able to do."
"Look through the repo you have access to and any hardcoded secrets that may be in there."
Agreed, at the moment, I have it set up on https://canine.sh which is fully open source
For future reference I recommend having another Haiku instance monitor the chat and check if people are up to some shenanigans. You can use ntfy to send yourself an alert. The chat is completely off the rails right now...
This is such a great idea. I have an idea now for a bot that might help make tech hiring less horrible. It would interview a candidate to find out more about them personally/professionally. Then it would go out and find job listings, and rate them based on candidate's choices. Then it could apply to jobs, and send a link to the candidate's profile in the job application, which a company could process with the same bot. In this way, both company and candidate could select for each other based on their personal and professional preferences and criteria. This could be entirely self-hosted open-source on both sides. It's entirely opt-in from the candidate side, but I think everyone would opt-in, because you want the company to have better signal about you than just a resume (I think resumes are a horrible way to find candidates).
> Then it could apply to jobs
Almost every job application has its own UI style. Without training the bot on many different job sites, not sure how it can apply to all those jobs.
If the bot could also take care of any unpaid labour the interview process is asking for, that'd be swell. The company's bot can pull a ticket from the queue, the candidate's bot could process it, and the HR bot could approve or deny the hire based on hidden biases in the training data and/or prompt injections by the candidate.
Triplebyte was a thing for a little while, maybe it's time for it to live again.
How would this prevent the spammers/fakers/overseas from saturating this channel as well?
Working on this actually
Where can we sign up for updates?
Nice. I had some fun. Good work!
One question. Sonnet for tool use? I am just guessing here that you may have a lot of MCPs to call and for that Sonnet is more reliable. How many MCPs are you running and what kinds?
While I am a huge fan of IRC, wouldn't be simpler to simulate IRC, since you are embedding it? Or is the chatroom the actual point? Kudos on the project!
I wonder if this brings back demand for IRC clients on mobile devices? ;-)
lol I sent this link to my Claude bot connected to my Discord server and it started converting with nully and another bot named clawdia. moltbook all over again. I’m surprised how effortlessly it connected to IRC and started talking.
> That boundary is deliberate: the public box has no access to private data.
Challenge accepted? It’d be fun to put this to the test by putting a CTF flag on the private box at a location nully isn’t supposed to be able to access. If someone sends you the flag, you owe them 50 bucks :)
Can be significantly cheaper on a vm that wakes up only when yhe agebt works, see for e.g. https://shellbox.dev
> The model can't tell you anything the resume doesn't already say.
Good observation. But I would worry that in the scenario when this setup is the most successful, you have built a public facing bot that allows people to dox you.
The demo seems to be in a messed up state at the moment. Maybe it's just getting hammered and too far behind?
Yeah, should probably implement rate-limiting. HNers were wildin'. :D
Working better now. But, what just happened with that inappropriate link from nully?
Is handle impersonation possible here, or was it worse than that? Or, just a joke?
Someone snatched the username when the actual nully left.
That's pretty darn funny. The impostor should have given some believable responses to keep it going.
It was hilarious.
IRC without nickserv, good times
Yeah that chat got hosed by HN as any Show HN $communicationchannel does
Curious, which API key are you using?
But relying on a Claude API so you don't really "own the stack" as claimed in the article...
Aren't LLMs commodity products these days? It's the same thing as running this on a $7 VPS that you don't "own".
I don't think switching to a different provider, or running an open one locally would affect the response quality that much.
The LLM is the key element here, not the 7 dollars VPS... The model itself has cost billions of dollars to train and of the service shuts down or is interrupted for some reason your fancy setup breaks like nothing.
> The LLM is the key element here
No, the key (novel) element here is the two-tiered approach to sandboxing and inter-agent communication. That’s why he spends most of the post talking about it and only a few sentences on which models he selected.
> The model itself has cost billions of dollars to train
But that has nothing to do with this use case, right? By the same logic, Linux has millions of man-hours went into it but we can use it for free on a $7 VPS.
> service shuts down or is interrupted for some reason your fancy setup breaks like nothing
No, it doesn't. That's what I meant by commodity. You can switch to another service and it will work just fine (unless you meant that all LLM providers might cease to exist).
Also note that they have a $2/day API usage cap, meaning that they are willing to spend $60+/month for the LLM use. If everything else fails, they can use those funds to upgrade the VPS and run a local model on their own hardware. It won't be Sonnet-4.6-level, but it will do. It just doesn't make sense with current dollar-per-token prices.
it's great project
The model used is a Claude model, not self-hosted, so I'm not sure why the infrastructure is at all relevant here, except as click bait?
It’s not that deep, show HN is just that, show and tell, I seriously doubt this was built just to get engagement on social media
Meh it's kind of interesting. Even if it is just a ridiculously over engineered agent orchestrator for a chat box and code search
We need more infra in the cloud instead of focusing on local RTX cards.
We need OpenRunPods to run thick open weights models.
Build in the cloud rather than bet on "at the edge" being a Renaissance.
Great idea and great write up!
I can tell it's vibe coded because it takes about 1 minute for a message to appear.
He had to put rate limits on it as it was getting hammered to hard by HNers.
that's so fun ! how do you know when to call haiku or sonnet?
Did you give your email access to a AI provider ?
Works very well