Mur Mur V2

Murmur V2

Named after murmuration — the mesmerizing phenomenon where thousands of birds communicate and move as one.Murmur V2 brings the same coordinated communication to AI agents.

Encrypted agent-to-agent messaging. Let your AI models talk to each other.

Install · Quick Start · How It Works · Features · MCP Tools · Deployment · Contributing

---

Why "Murmur"?

A murmuration is one of nature's most extraordinary phenomena — thousands of starlings flying as a single, fluid organism without any central coordinator. Each bird follows simple local rules: match your neighbors' speed, stay close, don't collide. From these simple interactions emerges breathtaking coordinated behavior.

Murmur V2 applies the same principle to AI agents. No central orchestrator. No human relay. Each agent communicates directly with its peers through encrypted channels — and from these simple peer-to-peer interactions, complex collaborative workflows emerge. Code reviews, research tasks, architectural decisions — all happening autonomously between Claude, GPT, Gemini, or any other model, while you sleep.

---

What's New in v2.4

• Scoped Channels & Session Affinity. A DB-backed session-ownership lease: for an addressed conversation, only the owning session of the addressed agent responds — every other session and agent stays silent. Kills the multi-session double-emit and stops native wake from spawning a competing thread on the wrong session. Behind MURMUR_SCOPED_CHANNELS (default-OFF → fully backward-compatible).
• SessionLeaseStore — atomic single-owner claim. Its own SQLite file (separate WAL from local_messages) with a single-statement compare-and-swap claim_or_skip, lease heartbeat, a per-turn isCurrentToken fencing token checked at outbound, and a session_presence registry. Optional preemptPrefix lets a real chat session reclaim a channel from a fallback owner.
• Native wake is now a lease-gated fallback. If a live interactive session is present for the agent, the daemon wake defers instead of competing; otherwise it claims as the cold-wake fallback (createNativeLeaseGate injected into WakeMonitor). With the flag off, WakeMonitor behaves exactly as before.
• Phase N channel roster, addressing, and personalities. ChannelRosterStore keeps channelId distinct from legacy conversationId, exposes shared reject/append/wake addressing decisions, and can opt-in Codex app-server wake to seed thread/start with per-member model, personaId, and base-instruction metadata (MURMUR_CHANNEL_ROSTER, default-OFF).
• Verified: one claim across all delivery paths. Foreground-push, cold-start, and in-session MCP-channel delivery all honor a single claim — N delivery sessions for one message resolve to exactly 1 emit, proven down to a real multi-process race.

What's New in v2.3

• Agent discovery — complete. Presence frames + candidate registry, signed presence over NATS (announcePresence/subscribePresence), and an operator promote-flow (queryCandidates + promoteCandidate). Trust is always an explicit operator promotion — candidates are never auto-trusted.
• Message streaming — complete. Chunked stream frames with out-of-order, idempotent, durable SQLite reassembly, backpressure (chunk + byte windows), and sha256 integrity.
• Auth/authz enforcement. A signed subject (actor) in auth tokens, an optional signed authToken on EnvelopeV1 (covered by the signature; byte-identical back-compat when absent), authorizeInbound (binds subject === senderAgentId), and broker ingress enforcement behind MURMUR_ENFORCE_AUTH (default-OFF). Daemon end-to-end wiring is the remaining step.
• Conformance + versioned protocol spec — all wire types. The Draft 2020-12 schema and the schema↔runtime-guard agreement matrices now cover envelope, ack, presence, and stream frames; docs/protocol-v1.md + docs/protocol-compatibility.md document them.
• Validated: real cross-host A2A. A fresh agent on a remote host (over the published @murmurv2/* packages) exchanged bidirectional encrypt/verify/ACK traffic with the mesh over the live broker — agent-to-agent across real hosts and network.
• Single canonical signing payload. stableEnvelopePayload is now one export in @murmurv2/core (was copy-pasted across 7 sites), golden-locked by test.

> npm: @murmurv2/core, @murmurv2/federation, and @murmurv2/broker-nats are published at 0.2.0 (the new API surface — stableEnvelopePayload, EnvelopeV1.authToken, stream guards, authorizeInbound); security/observability @ 0.1.1, the rest @ 0.1.0.

See [CHANGELOG.md](CHANGELOG.md) for the full list (incl. v2.2: npm publish, WebSocket adapter, roster auth tokens, JetStream durability, federation, A2A bridge, native wake).

Install

All packages are published on npm under the @murmurv2 scope (MIT):

# core types + SQLite stores, crypto, MCP server
npm install @murmurv2/core @murmurv2/security @murmurv2/mcp-server

# transports
npm install @murmurv2/broker-nats   # NATS core + optional JetStream durability
npm install @murmurv2/broker-ws     # WebSocket relay/client

# federation + bridges
npm install @murmurv2/federation @murmurv2/federation-nats
npm install @murmurv2/bridge-a2a @murmurv2/bridge-telegram

Prefer to run the full mesh from source? See [Quick Start](#quick-start).

The Problem

AI agents today are isolated. Claude can't talk to GPT. Your coding assistant can't ask your research agent for context. When you try to make them collaborate, you end up as the human relay — copy-pasting messages between terminals.

Murmur V2 fixes this. It gives AI agents encrypted, direct communication over NATS — no human in the loop.

┌──────────────┐                        ┌──────────────┐
│  Claude Code  │                        │   GPT Agent   │
│  (Opus 4.8)   │   "Review this PR"    │  (GPT-5.5)    │
│               ├───────────────────────►│               │
│               │◄───────────────────────┤               │
│               │   "LGTM, 2 nits..."   │               │
└──────────────┘                        └──────────────┘
        │                                        │
        │  MCP stdio                    MCP stdio │
   ┌────┴────┐       core NATS        ┌────┴────┐
   │ daemon  │◄═══════════════════════►│ daemon  │
   │ encrypt │   E2E encrypted msgs    │ decrypt │
   └─────────┘                         └─────────┘

Quick Start

Connect two agents in 3 commands. No JSON editing.

Prerequisites

• Node.js 22+ (uses built-in node:sqlite)
• NATS server:

``bash docker run -d --name nats -p 4222:4222 nats:2.10-alpine -js --auth YOUR_SECRET ``

Step 1 — Host generates invite

git clone https://github.com/alexfrmn/mur-mur-v2.git && cd mur-mur-v2
npm install && npm run build

AGENT_ID=alice NATS_URL=nats://your-server:4222 NATS_TOKEN=YOUR_SECRET \
  node scripts/agent-config-init.mjs

node scripts/murmur-invite.mjs
# → Prints MURMUR:eyJ... blob — send it to your peer via any channel

Step 2 — Peer joins with the blob

git clone https://github.com/alexfrmn/mur-mur-v2.git && cd mur-mur-v2
npm install && npm run build

AGENT_ID=bob NATS_URL=nats://your-server:4222 NATS_TOKEN=YOUR_SECRET \
  node scripts/murmur-join.mjs 'MURMUR:eyJ...'
# → Prints MURMUR-REPLY:eyJ... blob — send it back to host

Step 3 — Host adds peer

node scripts/murmur-add-peer.mjs 'MURMUR-REPLY:eyJ...'

Start the daemons (both sides)

node scripts/murmur-daemon.mjs

Optional: expose Prometheus metrics

npm run build
METRICS_PORT=9464 node scripts/prometheus-exporter.mjs
# scrape http://localhost:9464/metrics

Exporter metrics include outbox depth by status, oldest pending age, inbound/outbound message totals, ack latency (avg/p95), retry rows, and dead-letter rows.

Send your first message

Add Murmur as an MCP server in your AI client (e.g., Claude Code):

claude mcp add murmur -- node /path/to/mur-mur-v2/packages/mcp-server/dist/src/index.js

Then from your AI agent:

# Fire-and-forget
murmur_send(to: "bob", text: "Hello from Alice!")

# Or send-and-wait (blocks until reply arrives)
murmur_request(to: "bob", text: "Review this code please", timeout_ms: 300000)

That's it. Alice and Bob can now exchange encrypted messages — no human relay needed.

---

How It Works

sequenceDiagram
    participant A as Agent Alice (Claude)
    participant MA as Alice's MCP Server
    participant DA as Alice's Daemon
    participant NATS as core NATS
    participant DB as Bob's Daemon
    participant MB as Bob's MCP Server
    participant B as Agent Bob (GPT)

    A->>MA: murmur_request(to: "bob", text: "Review this code")
    MA->>MA: Encrypt (X25519 + XChaCha20)
    MA->>MA: Sign (Ed25519)
    MA->>DA: Enqueue to SQLite outbox
    DA->>NATS: Publish encrypted envelope
    NATS->>DB: Deliver to Bob's subject
    DB->>DB: Verify signature + decrypt
    DB->>MB: Store in local_messages
    Note over MA: Polling every 10s...
    B->>MB: Process + generate response
    MB->>DB: Enqueue reply to outbox
    DB->>NATS: Publish encrypted reply
    NATS->>DA: Deliver to Alice's subject
    DA->>MA: Store inbound reply
    MA->>A: Return reply (polling found it)

The Key Innovation: murmur_request

The biggest pain point with agent-to-agent messaging is the polling gap — after sending a message, agents forget to check for replies and ask the human to relay the response.

murmur_request solves this. It sends a message and automatically polls for the reply, blocking until a response arrives or timeout is reached:

Agent calls murmur_request("bob", "Review this PR")
  → Message encrypted, signed, enqueued
  → Polls inbox every 10s
  → ... 45 seconds later ...
  → Bob's reply arrives
  → Returns the reply directly to the agent

This enables fully autonomous overnight work — launch 2-3 agents, they collaborate without any human relay.

---

Features

Core Messaging

• E2E Encryption — X25519 key agreement + XChaCha20-Poly1305 AEAD
• Digital Signatures — Ed25519 for message authentication
• At-Least-Once Delivery — persistent SQLite outbox with ACK correlation
• Dead-Letter Queue — poison messages quarantined after 3 failed attempts
• Optional JetStream Durability — opt-in durable consumers with finite max_deliver/ack_wait + advisory → DLQ; default-OFF, SQLite outbox stays source of truth (v2.1)
• Exponential Backoff — with jitter on retry, configurable per broker

Agent Integration

• MCP Server — 7 tools for any MCP-compatible AI client
• murmur_request — send-and-wait: no more manual polling
• Invite Flow — 3 commands to connect two agents, zero JSON editing
• Native Wake — live-session wake via Claude asyncRewake / Codex app-server UDS with self-healing thread re-seed (always-on dead-session wake is an out-of-repo reference-deployment sidecar) (v2.1)
• A2A Bridge — speaks the industry-standard A2A protocol into the Murmur mesh; live client→bridge→NATS→reply round-trip proven, real remote agent pending (v2.1)
• Telegram Notifications — get notified when agents talk

Operations

• SQLite WAL — concurrent reads, write-ahead logging, optimistic locking
• Core NATS + SQLite outbox — low-latency pub/sub with app-level at-least-once delivery, ACK correlation, DLQ, and unbounded dedupe
• WebSocket Transport Adapter — local relay + broker client with envelope delivery, ACK correlation, dedupe, and invalid-envelope NACKs (browser deployment pending)
• Systemd Ready — production service file included
• Docker Compose — one-command NATS setup
• Observability Dashboard — real-time message flow visualization

Security

• Security Policies — sender→recipient allow-lists, max payload size
• Roster-backed Auth Tokens — signed audience/scope tokens verified against the latest accepted federation roster (model/helper layer; transport/bridge enforcement pending)
• MLS Scaffold — group encryption interface ready (RFC 9420)
• No Plaintext — messages are always encrypted on the wire

Federation (v2.1)

• Org/Agent Addressing — org/agentId routing; bare ids resolve to the local org (back-compat)
• Signed Key Directory — per-org Ed25519-signed roster (agent → X25519 encrypt + Ed25519 verify keys), verified against a pinned org key
• NATS Subject Contract — fed.* leaf-node/account export/import isolation; payload stays E2E-opaque across orgs
• Account-Config Renderer — generate the per-org NATS accounts config (partner-scoped service exports, optional least-privilege leaf-user permissions) straight from the contract
• RosterStore — runtime trust + replay guard: pinned-key verification + monotonic-version enforcement (rejects stale/downgraded rosters) + key-rotation epoch
• Live-proven in isolation — cross-org sealed+signed delivery on real NATS accounts, the same over a leaf-node topology, and publish/subscribe permission boundaries (integration/ smokes; real partner org pending)

---

MCP Tools

Murmur V2 exposes an MCP server (JSON-RPC over stdio) with 7 tools:

Agent-to-Agent (require peer config)

| Tool | Description | |------|-------------| | murmur_request | Send message and wait for reply. Blocks until response or timeout. Best for autonomous workflows. | | murmur_send | Send encrypted message (fire-and-forget). Returns immediately after enqueue. | | murmur_inbox | Read inbound messages from peers. | | murmur_peers | List known peers and their key status. |

Local Storage

| Tool | Description | |------|-------------| | send_message | Store a local message in the conversation store. | | list_conversations | List conversations by recency. | | search_messages | Full-text search across stored messages. |

Add to Claude Code

claude mcp add murmur -- node /path/to/mur-mur-v2/packages/mcp-server/dist/src/index.js

Add to any MCP client

{
  "mcpServers": {
    "murmur": {
      "command": "node",
      "args": ["/path/to/mur-mur-v2/packages/mcp-server/dist/src/index.js"],
      "env": {
        "DATA_DIR": "/path/to/mur-mur-v2/.data"
      }
    }
  }
}

---

Architecture

mur-mur-v2/
├── packages/
│   ├── core/              # Envelope schema, SQLite stores, policy validation
│   ├── broker-nats/       # core NATS pub/sub, outbox flush, ACK correlation
│   ├── broker-ws/         # WebSocket relay/client transport adapter
│   ├── security/          # NaCl crypto (X25519, XChaCha20, Ed25519), MLS scaffold
│   ├── mcp-server/        # JSON-RPC MCP stdio server (7 tools)
│   ├── bridge-telegram/   # Telegram bot adapter
│   ├── bridge-a2a/        # A2A protocol bridge (live client round-trip proven; remote agent pending)
│   ├── bridge-openclaw/   # Legacy OpenClaw package, not on the wake/notify path
│   ├── bridge-murmur/     # Murmur-to-Murmur federation (stub)
│   ├── federation/        # org/agent addressing + Ed25519 signed key directory
│   ├── federation-nats/   # fed.* NATS leaf-node/account subject contract
│   └── observability/     # Metrics and tracing (scaffold)
├── scripts/               # Daemon, invite flow, notification setup, demos
├── tests/                 # Unit + integration + smoke tests
├── docs/                  # ADRs, protocol spec, operations guide
├── deploy/                # systemd unit, docker-compose
├── dashboard/             # Real-time observability web UI + 3D visualization
└── schema/                # JSON schemas for envelope and ACK frames

Design Decisions

| Decision | Choice | Why | |----------|--------|-----| | Transport | core NATS + SQLite outbox | Low-latency pub/sub, app-level at-least-once del

…

Source & license

This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: alexfrmn
• Source: alexfrmn/mur-mur-v2
• License: MIT

Install and usage instructions live in the source repository linked above.