From Idea to Playable
AI-Assisted Game Development Bootcamp
Bring a game idea on Monday. Publish a playable v0.1 prototype by Friday.
5 days · 2 hours per day · 10 live mentorship hours · Monday–Friday
Galactic Omnivore Community Members attend for free.
Your Friday outcome
v0.1
A playable game.
A public URL.
A process you can repeat.
Your idea, playable
A practical five-day game development course where students turn their own game idea into a playable, published prototype. Learn a repeatable development workflow you can continue using independently: Godot, Codex, Git, GitHub and itch.io.
Create your own prototype and a working gameplay loop. Plan features with AI, review implementation plans, test generated changes, write professional bug reports, commit and push with Git, publish a playable v0.1, conduct an external playtest, and create a backlog for independent development.
Who this is for
People with a game idea but no idea where to start; beginner developers; designers and artists; students; people learning AI-assisted development; developers who struggle to finish prototypes; creators ready to publish and share playable builds. No professional programming experience is required.
What you will build
Your own small game: a working core loop, a clear goal, feedback, a beginning and an ending, and a restart. Every student may make a different game. Everyone follows the same development process.
How the course works
AI PROPOSES. CREATOR APPROVES. AI IMPLEMENTS. CREATOR VERIFIES.
IDEA → DEFINE → LIST FEATURES → AI PLAN → CREATOR APPROVAL → AI IMPLEMENT → RUN → OBSERVE → QA TICKET → FIX → VERIFY → COMMIT → PUBLISH → PLAYTEST → ITERATE
Ask AI to inspect the existing project and propose small, testable steps with dependencies, risks and acceptance criteria. Review the plan before implementation. If the scope changes materially, stop and approve the revised plan. Run and verify every change.
Five days. One repeatable process.
Day 1 — Idea to first playable v0.0.1
Turn an idea into something playable.
Idea vs prototype vs finished game. Define the player's main verb, goal and obstacle; scope v0.1. Set up Godot, Codex, Git and GitHub. Make the first AI interaction, player movement, interaction and gameplay mechanic. Make the first commit. Something can be controlled and something can be interacted with.
Day 2 — Directing AI v0.0.3
Learn how to develop with an AI coding agent.
Give AI a feature list. Ask it to inspect the project, identify systems, dependencies and risks, and propose small testable steps with acceptance criteria. Review and approve before implementation. If scope changes materially, stop and approve a revised plan. Run, test and report bugs using structured QA tickets. Implement small gameplay features: collectible, health, timer, enemy, score, power-up, inventory, projectile, dialogue or level transition. Finish with a recognizable gameplay loop.
QA ticket and AI debugging template
QA TICKET — BUG REPORT Subject Steps to Reproduce Expected Actual Still Working Evidence: screenshots / video, debug logs, console / error message Environment: game version, operating system, engine version, build type, Git commit AI debugging instructions: Inspect relevant code and identify the probable root cause before editing. Do not modify unrelated systems. Apply the smallest safe fix. Explain the probable root cause, list changed files, describe the change, give exact verification steps, and report remaining risks or unresolved issues.
Day 3 — Turn the prototype into a game v0.0.5
Can another person understand the game without an explanation?
VERB: what the player does. OBJECT: what they interact with. RULE: what happens. GOAL: why. FEEDBACK: how they understand. Add start and gameplay states, success, failure, restart, UI, visual feedback, appropriate audio and onboarding. Observe another person playing without explaining. Record Expected, Observed, Player confusion, Player behavior, Player feedback and Highest-priority change.
Day 4 — Ship it v0.0.8
Turn the local project into something other people can play.
Git workflow, commits, repositories, builds, export, CI/CD and automated deployment. Preferred pipeline: Godot → Git → GitHub → automated build → itch.io → playable URL. Publish a page with a title, one-sentence description, screenshots, how to play, controls, version, known issues, credits and feedback mechanism. Send the publicly playable build to external testers and join the cohort jam when available.
Day 5 — Independent creator v0.1.0
Continue without the mentor doing the work for you.
Independently understand an instructor's change request, list features, ask AI to plan, review and approve, implement, run, report and fix bugs, verify, commit, push and publish. Mentor intervention is limited to genuine blockers. Retrospective: WHAT EXISTS; WHAT WORKS; WHAT DOESN'T; WHAT PLAYERS SAID; WHAT I WOULD BUILD NEXT. Submit your playable itch.io URL and evidence of at least 10 reviews/ratings for human verification. The requirement stays at 10 even above 100 participants; verification may follow the live week.
Tools used
Godot · Codex · Git · GitHub · itch.io
Godot + Codex + GitHub + itch.io is the primary workflow. Git keeps each verified step recoverable.
What you will know after the course
How to turn an idea into a feasible scope, direct an AI coding agent, verify changes, report bugs, manage versions, publish a build, collect feedback and choose what to build next.
GO Prototype Creator — Level 1
Demonstrate Scope, Build, Debug, Version and Ship. Your instructor checks independence and evidence, not visual polish.
Publish a playable v0.1 on itch.io and submit evidence of at least 10 reviews/ratings. The minimum remains 10 even when a cohort has more than 100 people. Human verification can take place after the live week.
The success criterion: you can continue developing independently.
Upcoming cohorts
First cohort: internal GO team and existing Community members. Repeat attendance is welcome.
Loading dates and available places…
Standalone access is coming soon
No price has been announced. Join the price announcement waitlist for course availability and future video-bundle access. This does not reserve a cohort seat.
A prototype, and the skills to keep going
This is a practical system for turning game ideas into playable prototypes. It is not a complete Godot, Unity or s&box course, a programming degree, a commercial game in five days, or an AI-makes-everything-for-you course. The workflow can later transfer to Unity, s&box and other coding agents.
FAQ
Who can attend the first cohort?
Internal GO team members invited by an organizer and existing Community members. Choose GOHQ (8 seats) or online (unlimited). The course is in English, led by Ivan Kikerkov.
Will sessions be recorded?
Yes. Recordings are reviewed and approved before release. Recording access requires active membership. After membership expires, separate video-bundle purchase is planned; purchases are coming soon.
How do reviews work?
Publish on itch.io, gather at least 10 reviews/ratings and submit evidence. A human checks the evidence. The threshold remains 10 for every cohort size. Jam rating queues can help distribute playtests; this is GO's completion requirement, not an itch.io publishing restriction.
Can I attend again?
Yes. Repeat attendance in another cohort is welcome, subject to eligibility and in-person capacity.
How much does it cost?
Community members attend free. Standalone purchases are coming soon: join the price announcement waitlist. Future paid purchases will have a no-questions-asked refund policy; bank settlement timing varies.