This week, we'll show you how to connect Hermes Agent to Buzz and build an AI-native workspace around your agents.

The easiest way to picture Buzz is a Slack-style workspace where humans and AI agents work together.

Paired with Hermes, you can bring agents with memory, Skills, scheduled tasks, and approval workflows directly into the same workspace as your team.

By the end of this guide, you'll understand how Buzz works, the two best ways to connect Hermes, and an easy Skill that can handle most of the setup for you.

As always, after our main guide, stick around for the Midweek Edge for this week's AI news and Looking Ahead, where we compile the top AI research, trends, and workflows.

Let's get right into things!

Buzz Explained

Buzz is Block's open-source workspace where humans and AI agents share the same channels.

Think of it like Slack designed from the beginning for both people and agents.

You can create channels, send DMs, start threads, share files, react to messages, and work with agents in the same place. 

The only difference is that agents are the actual members of the workspace, with their own identity, channels, and jobs.

You can use a Block-hosted Buzz community to run it or run the relay on your own infrastructure.

Buzz is fully open source. Hermes is too.

Why Pair Buzz With Hermes

Hermes Agent is Nous Research's open-source agent with built-in memory, Skills, scheduled tasks, approval workflows, and support for different AI models.

It can run locally, on a VPS, or on other infrastructure you control.

Hermes already works through platforms like Telegram, Discord, Slack, and WhatsApp. When you pair Hermes with Buzz, you get a shared workspace designed around humans and agents working together.

With the full gateway connection, your existing Hermes agent can keep its memory, Skills, approvals, sessions, and scheduled tasks while working through Buzz channels and DMs.

For example, you could have:

→ A research agent dropping a morning report into your content channel

→ A coding agent working inside an engineering channel

→ An operations agent answering team questions from its saved knowledge

→ A scheduled Hermes task delivering a weekly report directly into Buzz

The combination becomes extremely useful. Hermes runs the agent, while Buzz gives that agent a workspace your entire team can work in.

Two Best Ways to Connect Hermes With Buzz

Hermes officially supports three different Buzz connection methods.

For most people, there are two worth knowing.

1. Run Hermes Inside Buzz Desktop

This is the fastest way to try the combination.

If Hermes is already installed on your computer, Buzz Desktop can detect it automatically.

Open:

Buzz Desktop → Settings → Runtimes

Hermes should appear as one of your available runtimes.

From there, create your Hermes agent and add it to the channels where you want it to work.

Buzz can also show the models you already have configured inside Hermes, so you can choose which model that agent should use for the session.

There is almost nothing else to configure.

Important: Keep Who can talk to this agent set to Owner only.

Buzz currently auto-approves Hermes tool permission requests through this Desktop connection, including terminal commands. Until that changes, do not open a local Hermes runtime to everyone in your workspace.

This route is best if you simply want to test Hermes inside Buzz before committing to a larger setup.

2. Run Buzz Through the Hermes Gateway

If Hermes is already part of your actual workflow, this is the setup we recommend.

Hermes stays running on your own machine or server, while Buzz is the other place you can talk to it.

This keeps the full Hermes setup intact:

→ Memory

→ Skills

→ Scheduled tasks

→ Approval workflows

→ Existing sessions

→ Your current Hermes configuration

Before connecting, you'll need a Buzz community, the Buzz CLI, and a dedicated Buzz identity for your agent.

Then run:

Select Buzz and follow the setup.

Once configured, start your gateway:

Your Hermes agent can now respond inside Buzz channels and DMs, receive files, reply in threads, and deliver scheduled tasks back into the workspace.

Which one should you use?

If you're testing Buzz for the first time, use Route One.

If you already rely on Hermes and want your team to work with the same persistent agent, use Route Two.

There is also a third buzz-acp relay bridge for hosted Hermes identities where Buzz manages the transport, but most users won't need it.

Buzz + Hermes Skills

There is an even easier way to set up the full connection.

A third-party repository called buzz-skills, built by tonbistudio, contains Hermes Skills specifically for working with Buzz.

The main one is hermes-in-buzz.

It walks Hermes through the gateway setup itself, including the Buzz CLI, agent identity, relay connection, access controls, configuration, and a live test to make sure messages work in both directions.

Installing It

Clone the repository:

Then copy the connection Skill into your Hermes Skills folder:

On Windows, your default Skills folder is:

Start a fresh Hermes session afterward so it detects the new Skill.

Then run:

Or simply tell Hermes:

Hermes will walk you through the rest.

The repository now includes two other useful Skills as well:

→ buzz-media-attachments helps Hermes send local videos, GIFs, and other media files directly into Buzz.

→ buzz-self-hosting helps you set up and troubleshoot your own Buzz relay.

You only need hermes-in-buzz for the basic connection.

One important note: These Skills are community-built and are not officially shipped by Nous Research.

The repository is, however, MIT-licensed, and the full source is available to inspect. We recommend reading it before installing, using a separate Buzz identity for your agent, and never pasting private keys or credentials into chat.

Our Recommendation

Start with one Hermes agent and one Buzz channel.

Use the Desktop route to see whether the workflow actually improves how you work.

If you find yourself using it every day, move that agent onto the Hermes gateway so its memory, Skills, approvals, and scheduled tasks stay intact.

Then start adding agents only when another clear job needs one.

Anthropic just released the much-anticipated Fable 5.1 model.

Full Fable 5.1 masterclass (complete overview, how to migrate, tips & more):

OpenAI will stop supplying its models to Cursor on November 12 following SpaceX's $60B acquisition of the company.

A federal judge ruled the Pentagon's “supply chain risk” blacklist of Anthropic was illegal and blocked the designation.

The court called the move unlawful retaliation after Anthropic refused certain military uses of Claude.

OpenAI says ChatGPT Ads reached a $1B annualized revenue run rate less than 200 days after launch.

The platform now has tens of thousands of advertisers, with self-serve access expanding across India, Europe, the Middle East, and North Africa.

Google just released Gemini Omni 1.1 Flash for AI video generation and editing.

→ Extend scenes up to 40 seconds
→ Set the first and last frames
→ Draft at 360p and upscale to 4K

Try it now in Google AI Studio or Google Flow.

Thomson Reuters launched Thomson, its own AI model built around 175 years of legal, tax, accounting, and news expertise.

The company says it spent $40M building the model, and its own benchmarks put Thomson competitively with Claude Opus 4.8.

Two models are getting a lot of attention right now:

→ Fable 5.1: New sightings reportedly show a claude-fable-5-1 identifier appearing through Amazon Bedrock, but Anthropic has not announced the model.

→ Grok 4.7: Elon Musk says initial training is complete, with xAI now adding a large amount of SpaceX company data before release. It is expected in early September.

Keep both on watch until official release details drop.

23 AI financial subsectors worth watching (for all the investors out there).

If you made it this far, thank you for reading, and we hope you found this week’s edition valuable.

If you enjoy reading, please forward our newsletter to someone you think would benefit from it.💙

Our promise to you: Every Wednesday, at 7 am EST, we’ll cut through the AI noise and send you human-curated AI content to make sure you stay ahead.

YouTube

Live Now: AI Agents: The Only Skill That Matters in 2026

X (Twitter)

Live Now: 8 CRACKED Claude Skills you NEED to install now

Free Skool Community

For exclusive behind-the-scenes, including AI prompts, cheatsheets, and PDF guides, join the free AI Edge Skool community here:

Interested in sponsorship or a partnership? Get in touch at [email protected]