Inside the Latest Pixels AMA: Open Source, Bounties, and a Live Zombie Demo

Yesterday's Pixels AMA started rough: a CDN migration to Cloudflare knocked the game offline right as it began, leaving Luke unable to demo anything for the first fifteen minutes. But once things stabilized, it turned into one of the most substantive open source walkthroughs the team has given: a live look at how community members can actually get code into the game, capped off with Luke building a working multiplayer zombie feature on stream in about twenty minutes using Claude Code.

Here's the full recap.

The State of the Game

Luke opened with some perspective on where Pixels is at two years post-TGE: the team has shipped 72 updates since launch, which is an unusually high cadence for a live game. The candid framing was that play-to-earn's first iteration was never meant to be permanent — the team has spent two years experimenting to find a version that's actually sustainable, learning what works and what doesn't along the way.

The Community Hub Is Now the Center of Gravity

The community.pixels.xyz hub, which went live just last week, is quickly becoming the official front door for contributing to Pixels — whether or not you can write code.

Two things stood out:

1. Requests: Anyone can submit an idea and the community votes on it. The top-voted request right now is for better industry search, industry timers, and improved mobile UI, things currently only available through third-party extensions.
2. Bounties: The Pixels team is attaching Pixel rewards directly to top requests (100,000 Pixel for the industry search/tracker request, for example). Luke confirmed this is drawing from token allocation that was already reserved for ecosystem rewards in the tokenomics, and called it possibly the best long-term use of that allocation. Notably, he also said the team plans to open up bounty-funding to anyone, not just the official team.

The first outside contributor to get a commit merged into the main game (a developer known as Mixer) has already received their bounty payout. Other developers, like kingrest, have merge requests in progress for features like the chat and friends system.

Two Paths to Building on Pixels

Luke laid out the two ways community members can currently build:

1. Reforge / Realms: a no-code tool for spinning up your own standalone version of Pixels. You get your own map, your own inventory system per realm, and a shareable link the moment you deploy. It's the lower-friction option, and it's the one the team recommends if you want to test an idea and get it in front of people fast, without touching a server.

2. Full open source client + server: the complete Pixels codebase, hosted on GitLab, which players can clone and run locally. This is a bigger lift (Node, Docker, and some command-line setup are required) but gives full creative control, including server-side logic that Realms alone can't touch.

Luke was explicit that if something built in a Realm turns out well, it's genuinely fast for the dev team to lift it directly into the main game — that's the intended pipeline, not just marketing language.

The Live Demo: Zombies in Terra Villa

The most striking part of the AMA was Luke setting up the open source repo from scratch, live, using Claude Code to troubleshoot along the way (including fixing a missing local database and some broken tutorial-skip logic). Once it was running, he asked Claude Code to prototype a "zombies" feature:

1. Zombie NPCs that spawn and chase the player around Terra Villa
2. Confirmed to run server-side, not client-side, verified by opening two browser sessions and watching the same zombies move independently for both players
3. A rough zombie-themed reskin of the Terra Villa map, generated by asking the AI to edit the map file directly (Luke noted this was messier than the NPC work and that map editing by hand is still the better approach for anything you actually want to ship)

The point wasn't that the feature was polished, it wasn't, but that a non-trivial, working multiplayer prototype came together in roughly twenty minutes without Luke opening the underlying code himself. He described this as the real shift open source unlocks: even non-technical community members can now get surprisingly far by directing an AI coding agent rather than writing code by hand.

Licensing, in Plain Terms

Luke confirmed the default license lets you do essentially anything with the codebase. The one line the team wants to hold: you can't fork the game and fully cut the Pixels ecosystem out of it: for example, launching a completely separate token off the codebase. Beyond that, building your own spinoffs, mini-games, or standalone servers is fair game.

Chubkins Update

Chubkins is around 90% feature-complete, with over 50,000 downloads. The remaining roadmap:

1. One more major feature push: subscriptions and stronger multiplayer hooks (currently, only a small share of players actually play with another user despite the game being designed as multiplayer)
2. A polish pass afterward: fixing the smaller confusing systems (skin tokens, dust, etc.) flagged repeatedly in app store reviews

The team's stated goal is a build that can support a profitable paid user-acquisition funnel, which is why they're holding a higher quality bar before scaling up marketing spend.

On Merchant Ships and Future Content

Asked about merchant ships, Luke didn't have a specific update queued up, but gave an honest explanation for why: every new type of "chase content" risks devaluing the rewards players already earned, and the actual fix isn't more items, it's new game loops. This is part of why open source matters strategically: it lets the community and the team experiment with new mechanics in parallel rather than only through official updates.

He also connected this to a related audience question about rare items becoming NFTs, floated as one possible way to let progress carry between the main game and community-built forks, alongside Pixel token balances, which already work across versions since they're on-chain.

What This Means for Contributors

The review cycle: the team evaluates open source merge requests roughly once a month and merges a handful of the best-reviewed ones. Luke noted contribution volume is still low enough that everything submitted gets reviewed, but expects that to change as more people get involved, which is prompting the team to think about scaling their review process.

For anyone experimenting with Pixels extensions or companion tools right now, the throughline from this AMA is clear: the team wants community-built features to graduate into the main game, they're funding that with real bounties, and — as the zombie demo showed — AI coding tools have lowered the bar for what a non-engineer can prototype enough that it's worth taking seriously.