Pal Server

pal-server forgejo

Notes

Project Page 1
  • Pal Server - Project Page project-pal-server

    Vision

    Stream Palworld from an AWS g4dn.xlarge GPU instance to any Moonlight client over Tailscale. Salt-managed, self-contained, single-repo. Start with one player (devopsphilosopher), prepare to scale.

    User Stories

    Key Story Note Role Success Metric
    solo-stream TBD Player (Lucas) Smooth 720p/60fps Palworld via Moonlight, <40ms latency

    Architecture

    POC phase - architecture docs live in-repo at docs/architecture.md. Formal arch notes TBD after POC validates.

    • AWS CLI launches g4dn.xlarge spot instance (Tesla T4)
    • Salt-call --local applies states: NVIDIA drivers, Tailscale, Sunshine, Steam
    • Moonlight connects over Tailscale WireGuard tunnel
    • NvFBC capture + NVENC encode (~2ms latency)

    Board

    board-pal-server

    Status

    Repo bootstrapped. PR #2 merged (docs + project structure). POC work next: launch EC2, write Salt states, validate streaming.

    Milestones

    None yet - POC in progress.

    Repos

    Repo Platform Role Status
    pal-server Forgejo Salt states, IaC, docs Active
Board 1