OpenHuman Is The Hermes Agent Killer?

Added: 2026-05-26

[00:00:00]We've recently seen Hermes agent gain momentum over openclaw, mainly because Hermes is positioned as a self-improving agent that grows with you over time, combining autonomous execution with a learning loop. But now there is a new agentic harness trying to steal Hermes's spotlight, and it's called Open Human.

[00:00:20]Open Human is a partially open-source human first desktop agent under the GPL3 license. It's designed to become the memory and the door for everything you do across your tools. It's built with Rust and Tari and it aims to close the gap between what AI models can technically do and what they can actually know about you. The idea is quite simple here. You choose open human if you want something private desktop native. It is easy to set up and capable of quickly understanding your life workflows as well as tools without much friction. If you want the best AI tools, workflows, and drops before everyone else, join my free newsletter with the link in the description below, which is completely free. What makes Open stand out is that it's not just an AI agent wrapper. It actually uses a local first memory tree that continuously pulls data from different tools whether that is Gmail, Slack, GitHub, whatever you're basically using on a daily basis that is going to drive it into a structured markdownbased memory system that is stored locally in SQL light instead of relying on blackbox vector memory. One really interesting part is the Obsidian style wiki system where you can actually read and edit the AI memory system yourself. It also has 118 plus integrations built within. You have automatic background syncing every 20 minutes. Model routing plus token compression for cheaper long context usage as well as browser and computer control and even a Google Meet agent that can actually join and transcribe meetings into memory. And if you are to compare open human to different terminal first agents like Hermes and OpenClaw, this is something that feels much more consumer focused cuz it has many of the integrations that you would use on a daily basis that your large language model can learn from. It's desktop native and it's designed around understanding your life and workflow quickly with minimal setups. Now, one thing that comes to my mind right away is if you're using proprietary models, they can definitely train off of your data. You definitely want to be careful with whatever you're connecting within open human. The best solution is using local models as that is going to let you keep any of your proprietary data locally rather than having it sent over to the model to be trained upon. Now to get started, it is super simple. You can easily install it with native packages which is recommended. You can install this for whatever operating system you have, Mac OS, Linux as well as Windows, which is great. But say if you want an easy approach, say if you're a beginner to coding, you can simply go over to the Tiny Human AI website from OpenHuman, you can click on download and install for whatever operating system you have.

[00:03:03]And this is essentially an installer that will set up Tiny Human onto your computer. After installing OpenHum and once you open it up, you're going to be then sent over to the onboard flow. This is where you can first select your runtime. I'm personally running this locally, but you have the option to run it on the cloud. But still it is highly recommend that you run it locally. Then you can click on continue and this will initialize open human. You have the option to then sign in with your Google account, Twitter or GitHub. I personally use a burner Gmail account. And then once I have signed in, I can then click on get started. Now I'm personally running the custom setup method. I'm not using any sort of credit system. I don't need to use a simple built-in security or any of this. So what we're going to be doing is continuing with custom. And now this is where we can configure our language model for text voice as well as oath. Meaning you can select different providers and integrations. Now what I'm going to be doing is keeping this as default but you can configure this so that you can bring in your own API provider. Then what we're going to be doing is selecting the voice system that we want. We're going to be setting this as the default voice which ships with manage ST and TTS that just works directly from open human. You can also configure this so that you can use uh 11 labs or even open AI whisperer. Then what we can also do is select our oath connections like Gmail, Slack as well as notion. We're going to be keeping this as default as well.

[00:04:32]And there you go. You're sent over to the main open human dashboard. Now, one thing I will say I do not like is how much they're trying to monetize this app. I obviously understand that you're trying to make something out of this, but still the popups and all this need to go. Regardless, this is the main control panel of open human. You can essentially ask your assistant to do anything. And this is where as you keep using it, it's going to ingest the memory that it retains from all your conversations as well as whenever you're speaking to the open human agent, it is going to then ingest it into the knowledge bank. Now you have the ability to also manage all of your connections.

[00:05:11]You can also use open human directly through these different channels like you would with openclaw. You have multiple different integrations that can help you with chat productivity tools and automations, social and platform.

[00:05:24]Essentially setting up all the automations directly through a unified application like open human which is desktop native. You have your intelligence page. This is your memory page and this is where you have the ability to manage your tree as well as your context. You can reset your memory which I wouldn't do but you can essentially get a summary tree after you have ingested all your knowledge. You can even add in knowledge by basically ingesting your folders or files which would be important if you're working on a coding project. And your subconscious mind is essentially where you can add in active tasks and you can even edit and play around with your memory. Task lists are essentially different sorts of tasks that will be appearing over here based off of your different sub agents that are running different sorts of tasks based off the prompts that you give it.

[00:06:13]And calls and dreams are two new features that they're going to be releasing afterwards. I would definitely not even play around with the rewards.

[00:06:19]That is something that I don't know why they even added. But regardless, this is the base structure of open human and just to showcase how you can set up local models or any other provider because right now it is going to use up the free tier or the free credits that they have within your chat as well as your human tab. So what you can do is go over to settings then you can click on advance and then once you are in advance you can click on AI configuration within large English model as well as voice you can configure your local models as well as your providers. So, let's actually start testing this out. This is where we're going to give it a market research task on basically an analyzing open human versus Hermes agent, open claw, cloud code, all of the different harnesses. And now what we're going to have it do is focus on a couple of different topics like memory systems, autonomy, integrations, and a couple other areas. And then after it identifies a couple of these strengths as well as these different requirements, what we can do is then have it work on sending over the PDF that it will create locally. And then now I have it connected to Gmail. So then it will send it over to Gmail as well. This is a simple automation that I have created, but you can even create schedule tasks locally and different automations with different integrations as well. Now, right now, it looks like it is asking for a couple things. Uh, I think it's cuz I'm using the free tier. So, I'm going to now switch it over to my own provider. So, it looks like it has finished creating the report contents.

[00:07:53]And this is where you use multiple tools to actually get the contents. And now it is working on creating the report. And by the way, when you are connecting any of these providers, just be safe because you are essentially syncing your memory from these providers into open human. So don't sync it if you are not comfortable with any of the contents being used through open human. They do state that they don't actually train off of it, but they're just measures that you would want to take with personal information.

[00:08:21]And you can see right now that this is one of the memory sources that I have being ingested, which is my Gmail contents. And you can even build a summary tree which will then be displayed directly in this section over here. And overall, this is a good way to manage all of your memory sources. And looks like the task is complete. This is the open market research report on all of these different harnesses, which is incredible cuz look how precisely it was able to get the job done. I'm using the GPT 5.5 and research right now within Open Human. And you can see that it did get the job done with the condensed source list at the bottom. So, this was pretty cool as to how it was able to do it and how fast it was in getting this job done. Now, let's actually go over to our Gmail to see if it did send it over.

[00:09:08]And just like that, we have the email sent within the sent category with the research paper that not research paper, but the report that it was able to conduct with multiple sources comparing all the different harnesses. If you like this video and would love to support the channel, you can consider donating to my channel through the super thanks option below. Or you can consider joining our private Discord where you can access multiple subscriptions to different AI tools for free on a monthly basis, plus daily AI news and exclusive content, plus a lot more. Now, I know this was a super simple prompt that was sent in, but still regardless, it did get the job done. And essentially what I really like about open human is the obsidian style readable memory system with automatic ingestion from multiple sources like Gmail so that it's able to improve as you use it more and more and get the live context that it needs so that when you need to understand something from any of these different connectors you can get the response that you're looking for and it is going to get you the most up-to-date answer based off the large nage model that you have connected that is going give you the information from all of the different sources that you have provided. This is also something that works differently than all of those other harnesses like Hermes or even OpenClaw. It's more of a unified UI that lets you manage everything through this simple UI. And this way you're able to even deploy automations like you saw and even do various sorts of things with this application. Now, I'll leave all the links that I use in today's video in the description below. This is something that you can easily get started with today. I personally still like using Hermes over open human but for someone who is looking for an easier uh deployment system as well as something that is easy to work with. Open human could be a easy alternative to get started with. But with that thought guys, I hope you enjoyed today's video and got some sort of value. I'll leave all the links that I use in today's video in the description below. Make sure you go ahead and take a look at the second channel, join the newsletter, join the Discord, follow me on Twitter, and lastly, make sure you guys subscribe, turn on notification bell, like this video, and please take a look at our previous videos so that you can stay up to date with the latest AI news.

[00:11:17]But with that thought, guys, have an amazing day, spread positivity, and I'll see you guys really shortly. He suffers.