Original post
Garry Tan#121
Garry Tan#121
Shubham Saboo@Saboo_Shubham_
This is HOW my Hermes and OpenClaw Agents share a single brain.
Hermes Agent is the orchestrator with OpenClaw agents as the squad.
9:32 PM · May 21, 2026 · 23.3K Views
Shannholmberg tests gBrain as shared memory layer for Hermes Agent squad recording 97.60% R@5 on LongMemEval_s benchmark
Garry Tan highlighted the multi-agent results as state of the art.
1091.2K1541.8K120.2K
Sentiment
Positive users praise GBrain's shared memory layer for making specialist Hermes agents useful via shared context, while negative users call it bloat, a single point of failure, or copied work.
Pos
73.2%
Neg
26.8%
43 comments with sentiment.
Cluster Engagement
Digg Deeper
No Digg Deeper questions have been answered for this story yet.
Posts from X
Most Activity
Most Activity
VIEWS66.4KBOOKMARKS1.2KLIKES714RETWEETS90REPLIES46
Shann³@shannholmberg
I've started experimenting with gBrain + Hermes Agent
it's a shared memory layer that sits underneath my Hermes Agent company. every specialist reads from the same brain before they do anything
the architecture I'm currently testing:
> inputs flow in: my ideas, strategy context, research, social signals, performance data > gBrain holds it all in typed folders: people/, companies/, concepts/, ideas/, media/, newsletter/, projects/, operations/ > the orchestrator (my main hermes agent) reads gBrain before every task and writes durable context back > specialists (research, socials, outbound, newsletter, engineering) get read-first access so they wake up with full context
the flow goes like this:
> 1. research agent → gBrain: enriches the brain with new findings > 2. gBrain → read-first context before any work > 3. orchestrator → gBrain: captures durable decisions > 4. tools → orchestrator uses fresh information from X, web, news > 5. orchestrator → I bring synthesized decision support
so don´t think about gBrain as an agent, it's the shared memory layer that lets a company of agents act like a team, with cross reference and a centralized brain
38dViews 66.4KLikes 714Bookmarks 1.2K
Garry Tan@garrytan
GBrain is SOTA for this kind of use case
Shann³@shannholmberg
I've started experimenting with gBrain + Hermes Agent
it's a shared memory layer that sits underneath my Hermes Agent company. every specialist reads from the same brain before they do anything
the architecture I'm currently testing:
> inputs flow in: my ideas, strategy context, research, social signals, performance data > gBrain holds it all in typed folders: people/, companies/, concepts/, ideas/, media/, newsletter/, projects/, operations/ > the orchestrator (my main hermes agent) reads gBrain before every task and writes durable context back > specialists (research, socials, outbound, newsletter, engineering) get read-first access so they wake up with full context
the flow goes like this:
> 1. research agent → gBrain: enriches the brain with new findings > 2. gBrain → read-first context before any work > 3. orchestrator → gBrain: captures durable decisions > 4. tools → orchestrator uses fresh information from X, web, news > 5. orchestrator → I bring synthesized decision support
so don´t think about gBrain as an agent, it's the shared memory layer that lets a company of agents act like a team, with cross reference and a centralized brain
38dViews 30.5KLikes 163Bookmarks 151
Shubham Saboo@Saboo_Shubham_
This memory layer is @garrytan amazing open-source project Gbrain: https://github.com/garrytan/gbrain
39dViews 1.7KLikes 8Bookmarks 17
Shann³@shannholmberg
gBrain repo if you want to test it out: https://github.com/garrytan/gbrain
38dViews 512Likes 3Bookmarks 13
Shubham Saboo@Saboo_Shubham_
@garrytan It is awesome (I use it as a shared brain for my Hermes and OpenClaw Agents)
38dViews 340Likes 6Bookmarks 4
Shubham Saboo@Saboo_Shubham_
@Curtitoo1 I have had my OpenClaw agents running for a while and they are all individual telegram bots that i ccan interact with when i like.
Hermes is the orchestrator with openclaw squad giving me the flex to keep my team as individual agents
39dViews 225Likes 3
Curtito@Curtitoo1
@Saboo_Shubham_ why not use hermes for everything (vs. openclaw) ?
39dViews 241Likes 5
Kairos Data Labs@KairosDataLabs
@Saboo_Shubham_ This is so cool.
39dViews 93Likes 4
Cecilia@ceciliahanif
@Saboo_Shubham_ What made you pick openclaw for execution and Hermes for orchestration when Hermes is defacto built to be multi agent and openclaw needs...help with it?
39dViews 250Likes 2
Shubham Saboo@Saboo_Shubham_
@shannholmberg I have seen this somewhere. Did you just copy my entire post without really understanding anything?
Insane 🤯
38dViews 75Likes 4
Hunter Gon@gonlenidefi
@Saboo_Shubham_ So Hermes calls the plays and OpenClaw runs them. Makes sense.
39dViews 121Likes 2
Joona 🧠Neurodivergent 🦋@The_Null_Path
@garrytan Gstack Baby 🫡
38dViews 10.2KLikes 2
Gargi@Gargi__Gupta
@Saboo_Shubham_ Super intersting.
39dViews 68Likes 2
Rugbist@rugbist_
@Saboo_Shubham_ so the lore is that ur agents have a shared braincell and hermes is the one driving
39dViews 156Likes 1
Saransh@SaranshSaboo
@Saboo_Shubham_ This is awesome 🤩
39dViews 146Likes 1
Alpha Batcher@alphabatcher
@Saboo_Shubham_ so many any information about Hermes and OpenClaw
literally banger
39dViews 144Likes 1
Kiran Mohan@KMohan40821
@Saboo_Shubham_ Just a basic question do you see any lines of code or work on it directly anymore. Or it's all through agents/skills? Do you apply a higher abstraction layer of verifiability?
39dViews 139Likes 1
Invincible@InvincibleEdge
@Saboo_Shubham_ squad dynamics and shared memory is the part these setups rarely get right
39dViews 137Likes 1
Kunaal Naik@KunaalNaik
@Saboo_Shubham_ Do you have a tutorial video use this ?
39dViews 76Likes 1
Lalo@LaloLoops
Hey thanks for sharing. Do you have the researcher constantly researching stuff on its own and feeding gbrain?
Or only when you ask it for something in particular?
If on its how does it know what to research for, based on what reasoning it might decide to research for something?
38dViews 229
Load more posts