mesh-join

Connect and coordinate multiple mesh-llm nodes into a distributed mesh.

3.1k|376|Updated Feb 11, 2026
One-click install
npx skills add https://github.com/Mesh-LLM/mesh-llm --skill mesh-join
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: mesh-join
Source: https://github.com/Mesh-LLM/mesh-llm/tree/main/.skills/mesh-join
Command: npx skills add https://github.com/Mesh-LLM/mesh-llm --skill mesh-join

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Orchestrates and connects multiple mesh-llm nodes into a cohesive mesh, enabling private deployments with invite tokens, joining public meshes via auto-discovery, and managing named or published meshes across diverse networks.

Core Features & Use Cases

  • Private meshes: create and join with invite tokens; publish for discovery (mesh-name) and cross-network collaboration.
  • Public mesh & client roles: auto-join best public mesh, API clients-only nodes, and dual-role deployments.
  • Networking and routing: NAT/firewall traversal, LAN mdns, iroh relay-assisted discovery, and unified OpenAI-compatible API routing on port 9337.
  • Verification and visibility: status endpoints to confirm mesh composition and model availability across peers.

Quick Start

Start two mesh-llm nodes and form a private mesh with an invite token or join the public mesh with --auto.

Frequently Asked Questions about mesh-join

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I connect distributed LLM nodes across different networks into a private mesh?

To connect distributed LLM nodes into a private mesh, start two mesh-llm instances and form the network using an invite token. This coordinates multiple nodes across diverse networks for private deployments and cross-network collaboration.

Can I join a public mesh automatically without manually configuring network discovery?

Yes, you can join a public mesh automatically without manual configuration by using the --auto flag. This feature auto-discovers and joins the best available public mesh, supporting API client-only nodes and dual-role deployments.

Does mesh-join handle NAT and firewall traversal for local area network discovery?

Yes, mesh-join handles NAT and firewall traversal using iroh relay-assisted discovery and LAN mdns. This ensures mesh-llm nodes can connect and coordinate across diverse networks even behind restrictive firewalls.

How do I verify connected peers and model availability in a distributed mesh?

You verify connected peers and model availability in a distributed mesh by querying the provided status endpoints. These endpoints confirm the mesh composition and ensure model routing is accessible across all connected nodes.

What is the best way to route OpenAI-compatible API requests across multiple LLM nodes?

The best way to route OpenAI-compatible API requests across multiple LLM nodes is through the uniform API on port 9337. This unified interface handles model routing across the connected mesh-llm peers automatically.

When should I use a client-only node deployment instead of a full mesh setup?

You should use a client-only node deployment when you want to consume models from a public mesh without publishing your own resources. This dual-role capability allows lightweight API access without contributing compute power to the distributed network.