Humalike x Hermes - Social intelligence plugin for Hermes Agent
by•
One command gives your Hermes agent social intelligence. It decides when to speak, adapts to your group's tone and remembers who said what. Works in group chats on Slack, Telegram and WhatsApp.
Replies
jared.so
Hey!! Maks here, Founding Engineer at Humalike.
All you need to give your agent social intelligence is one command, that's it. First credits are on us :))
It works in group chats on Slack, Telegram, Discord and WhatsApp.
Feel free to ask any questions. I'm here to help :))
Fedica
Congrats @mcarmonas & team. What was the main challenge you wanted to solve when building Humalike?
Humalike
@hamza_afzal_butt tysm! Social skills & proactiveness!
A suggestion from someone who works with voice agents a lot: it would be great to see a turn-taking or interruption handling module exposed through the API. Right now most stacks handle barge-in in a pretty mechanical way, so a behavioral layer that knows when to pause, yield, or gently hold the floor would be a really meaningful addition.
@feride178675 Yep, exactly! That's what we mean by behavioral layer, and turn-taking API is acutally one of the 7 APIs we have available, although it's not tailor-made for voice. If you have use-case for voice where you need humanlike turn-taking, let's connect and chat:))
Humalike
@feride178675 :)))))
Love the concept. One question: can developers customize how talkative the agent should be depending on the workspace or group?
Humalike
@nayan_joshi Thanks! Yes! Talkativeness is part of the agent's personality, and each workspace/group runs its own, so the same agent can be all business in one room and playful in another.
And even without touching anything, thanks to social learning API, it reads the room and matches each group's rhythm on its own: quiet team, quieter agent.
Humalike
@nayan_joshi tysm!
The "decides when to speak" gating is the actually hard part here — an agent that stays quiet 90% of the time is worth more than one that answers every message, but that means the plugin holds group context the base Hermes agent doesn t. Where does the "who said what" memory live: inside my own Hermes runtime, or a hosted Humalike service the group messages get routed through? And is the speak/stay-silent decision a model call on every message (latency + cost per turn) or a lightweight local classifier?
@noctis06 Hey Noctis, these parts live on Humalike side. Both memory and classifier are on our side for few reasons:
1. ease of integration
2. easy way to ship improvements
3. the Hermes plugin uses Humalike APIs that are transferable from product to product. That means if you ever wanted to create your own agent you can easily use Humalike APIs and get similar behavior as Hermes x Humalike
Cost per turn is low from what we tested, when you sign up you get 10.000 free credits which should be enough for months of normal usage for hermes agent.
Let me know if you are concerned about something else specifically:))
What I'd want to know is how long the decision takes. In a live group the window where a message is still useful is a couple of seconds wide, and every should-I-speak check we tried added another model call, so by the time it came back yes the thread had moved on two messages and the reply read as odd. Is the gating a small fast classifier sitting in front of the main model, or does the same model produce both the decision and the message?
@dipankar_sarkar Hey, I think these are two separate things:
1. actual latency for making a decision
2. "group has moved on" problem
We solve 2 really well by handling interruptions and making sure agent doesn't respond to old messages. About 1, yes, it adds latency, and it's not perfect. In our testing it's not a deal breaker in chat, LLMs are fast enough (compared to humans that acutally have to type each message). I encourage you to give it a try and see how you like it, we would love your feedback!
I appreciate that this tackles a problem many developers notice after deployment. Technical accuracy alone rarely creates comfortable interactions. Will there be tools for testing personality consistency before releasing agents?
@john_michael31 Hey, yes! We have an audit page where you can paste transcript of any agent (powered by Humalike or not) and you will have detailed analysis of how it performs.
We also have CLI for it if you want to run this analysis on recurring basis e.g. once per week!
Humalike
@john_michael31 :))) https://humalike.ai/audit
Does the memory work across different group chats or is it per-group? Curious how much conversation history it needs before it starts picking up tone shifts.
@talhakhalidmtk One memory for all of your groups, platforms (discord, telegram, slack, whatsapp...) and threads. Give it a try and push it to the limit
TapRefer
so you're giving soul to hermes??
@jiteshghanchi Hahah, actually good that you ask, we should clarify.
We add whole new layer for turn-taking: decide when to speak vs stay stilent, that works well in group chats. We change how hermes acts when he is interrupted. We change the way he responds (short natural burst messages instead of long report-like messages). This is engineering work, not just a prompt.
We add Social Memory -> another layer of memory focused on people, their preferences, humor and opinions.
We add social learning that fires every few messages -> Hermes picks up new jokes, style and norms without you having to touch soul[.]md at all. It happens automatically many times during one chat.
Annnddd yes, we also improve your Soul based on our experience building 4 different humanlike agent products:))
You should give it a try, the jump you can make with Soul alone is 1.5x, with everything we got it's 10x
Humalike
@jiteshghanchi 👀👀👀
I want to know whether the hardest part was making the agent more human, or making it less eager to participate.
Humalike
@reda_roqai_chaoui Hahaha that's a good one! It depends a lot...
We built a AI Community Manager in the past that, if it was properly guided (through channels) and with a lot of constraints (behavior), the hardest part was how to make it human (infinite issues). Now, if you are building something with a lot of constraints and it's against product nature (having constraints is bad), making it less eager to participate is so hard.