mirror of
https://github.com/Mesh-LLM/mesh-llm.git
synced 2026-08-08 22:23:19 -04:00
Repo was transferred to the Mesh-LLM org. Update all GitHub URLs, install script defaults, auto-update endpoints, CLI defaults, UI links, docs site links, and container image references. Intentionally left alone: - michaelneale/llama.cpp (fork not transferred) - relay.michaelneale.mesh-llm.iroh.link (iroh relay domain, not a GitHub ref) |
||
|---|---|---|
| .. | ||
| console | ||
| Dockerfile | ||
| README.md | ||
mesh-llm Fly.io Console
Fly app running mesh-llm in --client mode — no GPU, just QUIC tunnels to mesh nodes.
| App | URL | Fly config |
|---|---|---|
| console | mesh-llm-console.fly.dev | fly/console/fly.toml |
Also available at www.mesh-llm.com and www.anarchai.org.
Architecture
┌─────────────────────────┐
Browser/curl ──HTTPS──→ Fly │ mesh-llm --client │
│ discovers mesh via │──QUIC──→ GPU nodes
│ Nostr, tunnels requests │
└─────────────────────────┘
Exposes :3131 (dashboard, chat, topology) and proxies inference to mesh GPU nodes.
Deploy
From the repo root:
fly deploy --config fly/console/fly.toml --dockerfile fly/Dockerfile
Run locally
# Same as what the Fly app runs — no Docker needed
mesh-llm --client --auto
Docker (local)
docker build -f fly/Dockerfile -t mesh-llm-console .
docker run -p 3131:3131 -p 9337:9337 mesh-llm-console