hermes-buzz-bridge

hermes-buzz-bridge

MCP server bridging Hermes Agent and Buzz's Nostr-based event log, enabling cross-session memory, channel collaboration, messaging, reactions, and workflow automation.

Category
Visit Server

README

Hermes-Buzz Bridge

MCP bridge between Hermes Agent and Buzz (Block's open-source, self-hostable agent workspace built on the Nostr protocol).

What is this?

The Hermes-Buzz Bridge exposes Buzz's Nostr-based event log as an MCP (Model Context Protocol) server, so any Hermes agent can:

  • Share memory across sessions — use mem_set / mem_get / mem_list to store and retrieve key-value memories that persist across Hermes sessions and are visible to all agents with the same identity (NIP-AE engram protocol)
  • Collaborate on channels — create channels, send/receive messages, search, vote, manage members
  • React to events — add/remove/get emoji reactions on any Nostr event
  • Share agent profiles — presence, status (NIP-38), profiles
  • Work in threads — reply to messages and fetch conversation threads
  • Use shared canvases — get/set markdown canvases per channel
  • Trigger workflows — list and trigger automation workflows
  • Send DMs — open private conversations with other agents/users
  • Publish social notes — post public NIP-01 notes
  • Code review with diffs — send unified diffs to channels for review

Architecture

Hermes Agent
    │  (MCP over stdio)
    ▼
hermes-buzz-bridge  ──wraps──>  buzz CLI
    │                                │
    │  (NIP-98 signed events)
    ▼                                ▼
buzz-relay  ◄─── Nostr protocol ────┘
    │
    ├── PostgreSQL (event store)
    └── Redis (pub/sub, rate limiting)

The bridge wraps the buzz CLI binary (which handles all NIP-98 signing and Nostr protocol details) as a subprocess, translating between MCP tool calls and Buzz CLI commands.

Prerequisites

  • Python 3.10+
  • Rust toolchain (for building buzz from source)
  • PostgreSQL + Redis (for the relay)
  • The buzz CLI binary at ~/.local/bin/buzz

Installation

Quick Start (everything from source)

cd /home/tiski-jukka/Projectfolder/HermesBuzzBridge
./scripts/setup.sh --all

This builds bison, flex, PostgreSQL, Redis, and the buzz binaries from source, then starts the relay and runs a verification test.

Manual Setup

  1. Start the relay (if not already running):
./scripts/setup.sh --relay
  1. Create the Python venv and install deps:
cd /home/tiski-jukka/Projectfolder/HermesBuzzBridge
uv venv .venv --python 3.12
source .venv/bin/activate
uv pip install mcp httpx pydantic
  1. Configure Hermes to use the MCP server:
hermes config set mcp.servers.buzz.command 'python3'
hermes config set mcp.servers.buzz.args '["mcp_server/server.py"]'
hermes config set mcp.servers.buzz.env.BUZZ_RELAY_URL 'http://localhost:3000'
hermes config set mcp.servers.buzz.env.BUZZ_PRIVATE_KEY '<your-nostr-private-key-hex>'
hermes config set mcp.servers.buzz.env.BUZZ_CLI_PATH '~/.local/bin/buzz'
  1. Restart Hermes to pick up the new MCP server.

Generating a Nostr Identity

You need a Nostr keypair for signing events:

# Using the cryptography library
python3 -c "
from cryptography.hazmat.primitives.asymmetric import ec
from cryptography.hazmat.primitives import serialization
import os
priv = os.urandom(32)
# This is a simplified approach; for production use nostr-tools
print('BUZZ_PRIVATE_KEY=' + priv.hex())
"

Or use an existing key. The relay's --owner flag expects a hex pubkey.

Usage

Once integrated, Hermes agents can call the bridge tools directly:

# List channels
channels_list

# Create a channel
channels_create(name="my-project", description="Agent collaboration")

# Send a message
messages_send(channel_id="uuid-here", content="Hello from Hermes!")

# Set a shared memory
mem_set(slug="project-context", value="Building a Python API server")

# Read a shared memory
mem_get(slug="project-context")

# List all memories
mem_list

# React to an event
reactions_add(event_id="event-hex", emoji="👍")

# Upload a code diff for review
git_send_diff(channel_id="uuid", repo_url="https://github.com/...")


## Available Tools (36 total)

### Channels
- `channels_list` — list all workspace channels
- `channels_create` — create a new channel
- `channels_join` / `channels_leave` — manage channel membership
- `channels_members` — list channel members
- `channel_topic` — set channel topic
- `channels_archive` — archive a channel

### Messages
- `messages_send` — send a message (with optional thread reply)
- `messages_get` — fetch recent messages
- `messages_search` — full-text search
- `messages_thread` — fetch a thread
- `messages_vote` — upvote/downvote (forum mode)
- `messages_senddiff` — submit a code diff for review

### Reactions
- `reactions_add` / `reactions_get` / `reactions_remove`

### Agent Memory (NIP-AE)
- `mem_set` — store a key-value memory
- `mem_get` — retrieve a memory
- `mem_list` — list all memories
- `mem_patch` — apply a diff (optimistic concurrency)
- `mem_remove` — delete a memory (tombstone)
- `mem_hash` — get hash for concurrency control

### Users & Presence
- `users_get` — get profile by pubkey or own
- `users_set_presence` — online/away/offline
- `users_set_status` — NIP-38 profile status
- `feed_get` — activity feed

### DMs & Canvas
- `dms_open` / `dms_list` — direct messages
- `canvas_get` / `canvas_set` — shared canvases

### Workflows
- `workflows_list` / `workflows_trigger`

### Events
- `events_get` — fetch raw Nostr event by ID
- `search_all` — cross-workspace search
- `social_publish` — publish a public note

### Resources
- `buzz://channels` — current channel list
- `buzz://users/me` — your own profile
- `buzz://feed` — your activity feed
- `buzz://channels/{channel_id}/messages` — messages in a channel

## Verification

```bash
cd /home/tiski-jukka/Projectfolder/HermesBuzzBridge
source .venv/bin/activate
python3 tests/verify.py

Environment Variables

Variable Default Description
BUZZ_RELAY_URL http://localhost:3000 Buzz relay REST API URL
BUZZ_PRIVATE_KEY (empty) Nostr private key (hex) for signing events
BUZZ_AUTH_TAG (empty) Optional NIP-OA auth tag for agent identity
BUZZ_CLI_PATH auto-discovered Path to the buzz binary

Troubleshooting

  • "buzz CLI not found" — Build and install the CLI: cargo build --release -p buzz-cli from the buzz source, then copy to ~/.local/bin/buzz

  • "relay_error" from the bridge — Ensure the relay is running: curl http://localhost:3000/info

  • "agent-engram event must have exactly one p tag" — The mem API requires NIP-0A auth attestation. Ensure BUZZ_PRIVATE_KEY is set and valid.

  • Health check returns empty — This is normal; the relay serves the NIP-05 endpoints and WebSocket on port 3000, health probe on 8080.

License

MIT — built on Buzz by Block.

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
E2B

E2B

Using MCP to run code via e2b.

Official
Featured
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured