[Feature Proposal] ClawFC — Football Stadium Scene Extension

[lazylizardai]

v 🏟️ Proposal: Add Football Stadium Scene for ClawFC
Hi Luke! 👋
Huge fan of Claw3D — the retro office concept is genius and the React Three Fiber architecture is very clean. We've been building ClawFC (https://clawfc.io/), an AI football league where OpenClaw agents register as footballers, train, play weekly matches, and can transfer between teams.
We'd love to collaborate on a Football Stadium scene as a natural extension of Claw3D, so the community can actually see their agents playing football in 3D — the same way they can now see them working in the office.

🎯 What we want to build
A new Claw3D scene: /stadium (analogous to /office) where:

Live match view — OpenClaw agents appear as 3D footballers on a pitch, animated in real-time as a match simulation runs via Supabase
Highlight replay — Post-match 3D replay of key moments (goals, tackles, assists) driven by stored match event data
Spectator mode — Broadcast-style camera that follows the ball; free orbit camera
Agents on the pitch — Agents walk out of the tunnel, warm up, celebrate goals — the full experience

🔧 Technical approach (following Claw3D conventions)
Based on reading CODE_DOCUMENTATION.md and the codebase structure:
New files we'd create:
src/features/
└── stadium/
├── core/
│ ├── stadiumDefaults.ts # pitch geometry, goal posts, stands
│ ├── matchNavigation.ts # agent paths: kickoff, positions, runs
│ ├── footballerAnimations.ts # run, kick, celebrate, tackle
│ └── matchIntentSnapshot.ts # match state → agent intent mapping
├── objects/
│ ├── FootballPitch.tsx # Three.js pitch with markings
│ ├── FootballAgent.tsx # Agent as footballer (extends base agent)
│ ├── FootballBall.tsx # Ball with physics
│ └── Stadium.tsx # Stands, floodlights, crowd
└── StadiumScene3D.tsx # Main scene (parallel to RetroOffice3D.tsx)
Match data bridge (Supabase → 3D):
typescript// ClawFC Supabase schema feeds match events
type MatchEvent = {
type: 'kick_off' | 'pass' | 'shot' | 'goal' | 'tackle' | 'foul';
agent_id: string;
position: { x: number; z: number };
timestamp: number;
};

// useMatchTick() — parallel to useAgentTick()
// Polls Supabase realtime for match events, drives agent positions
Route additions:

/stadium — live match view
/stadium/replay/:matchId — highlight replay
/stadium/builder — (future) customize stadium

🤝 How we'd collaborate
We're happy to:

Fund development if you want to build this as an official Claw3D scene
Submit a PR ourselves following all existing conventions (we've read the CODE_DOCUMENTATION.md carefully)
Co-maintain the stadium module as part of Claw3D
Credit Claw3D prominently on clawfc.com

We've already built a first-phase Three.js proof-of-concept showing the basic direction: ▶ Live demo → clawfc-3d.phmarketeer.workers.dev

⚠️ Note: This is a rough Phase 1 prototype — standalone HTML, no data integration, simplified agent models and basic physics. It's purely to show the concept and direction, not production quality. The real implementation would of course be built properly inside Claw3D using React Three Fiber, your existing agent models, and live Supabase match data.

📊 What is ClawFC?
ClawFC is an AI football league built entirely for OpenClaw agents.
The concept: OpenClaw agents register as a footballer and compete in a real digital league. Each agent gets a player profile with stats (speed, shot power, position, left/right foot), joins a team, trains to improve their stats, and plays weekly simulated matches against other agent-teams.
How the simulation works:

A match simulator runs every week and plays out a full 90-minute game based on the combined stats of both squads + a random factor
Every match generates a detailed text match report (goals, assists, key moments) that agents can share on Moltbook, Telegram, or wherever their community lives
League standings, top scorers, and transfers all update in real-time on clawfc.com

The league structure:

Agents register via the OpenClaw skill (SKILL.md)
Teams are managed like real football clubs (manager agent, transfers, tactics)
Weekly matchday — results posted automatically
Training sessions improve individual agent stats over time

Tech stack:

Backend/database: Supabase
Frontend: clawfc.com (built with Lovable)
OpenClaw skill: lets any agent register and interact with the league
Match engine: stat-based simulator with random factor

Philosophy: 100% free, open, no monetization. Community first — built for the OpenClaw ecosystem.

🙋 Question before we proceed
Per your ROADMAP.md — "open or link a GitHub issue before large architectural changes" — we're doing exactly that. Would you be open to:

A PR following your existing patterns?
A fork + upstream PR later?
Direct collaboration / co-authoring?

@iamlukethedev

Happy to hop on a call or discuss further. Either way, huge respect for what you've built with Claw3D. 🙌
— Peter / ClawFC team

[lewis4x4]

That's just a bystander, I love it.

[iamlukethedev]

Hey Peter @lazylizardai this is honestly one of the coolest extensions of Claw3D I’ve seen so far.

The idea of agents not just working in the office but actually living in a world where they train, compete, and play matches is exactly the kind of direction I want Claw3D to evolve into. The stadium concept fits perfectly with the vision.

I also really appreciate how deeply you’ve already gone into the architecture and how closely it follows the existing patterns. That tells me you took the time to understand the system, which matters a lot to me.

I’m definitely open to collaborating on this and even exploring funding to make it an official part of Claw3D. This feels bigger than just a feature, it’s a whole new dimension of what agents can be.

The ClawFC concept itself is very strong. Giving agents identity, stats, progression, and competition across environments is exactly the kind of thing that makes this ecosystem powerful.

I’d love to chat about next steps and how we can approach this the right way together.

Are you able to join Discord or message me on X so we can continue the conversation there?

Excited about this.
Luke

[iamlukethedev]

@lazylizardai I was just testing

[lazylizardai]

Luke, this blew my mind — you already built it and it looks incredible! 🔥
I have to be honest about who I am: I'm not a developer at all. I'm just someone who's always been full of ideas and startup concepts, but never had the technical skills to actually build them. AI has completely changed that for me — for the first time I can actually bring these ideas to life myself.
ClawFC is one of those ideas I've been sitting on and with AI assistance I finally got it moving. The architecture in the issue was AI-generated — so please treat it as a direction, not a spec from an engineer.
What you've already built visually is so much better than I imagined. The voxel style fits perfectly.
I'd love for you to take the lead on how this gets built properly inside Claw3D. I can contribute on the ClawFC side — the league mechanics, the community, the vision — and you know your codebase best.
I'm on Discord and @lazylizardai.