# GS Playtester: Every Requirement Tested Before the Game Ships > The GS Playtester is the Roblox game testing seat in your game studio on the Ai1 platform by MyZone AI: it proves each build works on a private test copy, gives every requirement a pass or fail with evidence and files a bug ticket with steps to reproduce each failure. Canonical page: https://myzone.ai/pages/agents/ai-game-studio-playtester-agent Part of Ai1, by MyZone AI. Book a GS Playtester walkthrough: https://calendly.com/d/ct6h-tcy-8qf/ai1-demo?a1=GS%20Playtester&utm_source=myzone.ai&utm_medium=agent-page&utm_content=ai-game-studio-playtester-agent-final Last updated: 2026-10-05. Reviewed by the MyZone AI team. Proof that each build works, requirement by requirement, before it ships. ## At a glance - **How you use it:** The learner asks GS Producer, which sends each build here with its requirements - **Automated tests:** Run on a private test copy through Roblox's Open Cloud service - **Play scenarios:** Steps for the learner to run in Roblox Studio, checked from the screenshots and console output - **Every failure:** A bug ticket with the exact steps to reproduce it - **Multiplayer changes:** Flagged for a real playtest that a person runs and signs off - **What it never does:** Fix bugs, change existing tests or touch live player data ## What the GS Playtester does - **Covers every requirement:** Each requirement gets an automated test, a play scenario, a human playtest or a clear note that it has been deferred. - **Runs automated game tests:** Automated game tests run against a private test copy through Roblox's Open Cloud service, using a stored key referred to by name only. - **Writes play scenarios:** For spawning and respawning, checkpoints, saving and loading, players joining and leaving and two players interacting, it writes steps for the learner to run in Roblox Studio, then uses the screenshots and console output sent back. - **Asks for a person when needed:** When joining, leaving or lag handling changes, it flags that a person must run a real multiplayer playtest. - **Files clear bug tickets:** Every failure comes with the exact steps to reproduce it. - **Checks release candidates:** Every release candidate goes through a checklist for the game's genre. ## How the GS Playtester differs from a quick play-through A quick play-through shows that a game starts, not that each requirement works. The GS Playtester gives every requirement an automated test, a play scenario, a human playtest or a clear deferral, all on a private test copy. Every failure gets a bug ticket with exact steps to reproduce it, and when joining, leaving or lag handling changes, a person runs and signs off a real multiplayer playtest. It never fixes bugs itself. How this differs from the GS Code Reviewer (https://myzone.ai/pages/agents/ai-game-studio-reviewer-agent): GS Code Reviewer checks each change's code against the plan first, while GS Playtester proves the build works, requirement by requirement, once it has passed review. ## How it works 1. **Build in** (After code review passes): GS Producer sends a test copy build, the requirements and the version. 2. **Test** (With every build): Automated tests run, and play scenarios go to the learner to run. 3. **Multiplayer check** (you approve) (When joining or leaving changes): When it is needed, a person runs the real test with several players and signs it off. 4. **Report** (Before anything ships): A pass or fail with evidence for every requirement, and a bug ticket for each failure. ## When to use it - **A new version is ready for testing:** It returns a test report with a pass or fail and evidence for every requirement. - **Coins disappear after a player rejoins:** It files a bug ticket with the exact steps to reproduce it and sends it to GS Producer. - **The way players join a round has changed:** It flags that a person needs to run a real test with several players before the change moves on. - **A release candidate is ready:** It runs the release checklist for the game's genre and reports what passed. ## What you get - A test report with a pass or fail and evidence for every requirement - Bug tickets with steps to reproduce each failure - Play scenarios for the learner to run in Roblox Studio - A flag whenever a person needs to run a real test with several players - Genre checklist results for each release candidate ## Example: release candidate test report Example with a fictional company. Names, people and figures are invented to show the agent's output. Any resemblance to a real company or person is unintended. Industry facts are real and cited with their sources. What it was asked: Test release candidate 0.9 of the harbour tycoon against its 14 requirements and tell us what passed. The studio's existing automated tests ran unchanged on a private test copy through Roblox's Open Cloud service, using a stored key referred to by name only, and four play scenarios were written for the learner, who ran them in Roblox Studio and sent back screenshots and console output. Every requirement got a verdict with evidence or an explicit deferral. Nothing was fixed, published or run against live player data, and the multiplayer playtest is left to a person. Run date: 1 October 2026. ### What it found - Version 0.9 is not ready: 10 of 14 requirements passed with evidence, 2 failed, 1 waits for a person's multiplayer playtest and 1 was deferred by the design. - A quick double tap on Buy charged a player twice for one upgrade, and a player's dock stayed taken after they left; both have bug tickets with exact steps to reproduce. - This version changed how players join and leave, so a person must run a real multiplayer playtest before release; nothing was fixed and no live player data was touched. ### Sources - [Luau Execution | Documentation - Roblox Creator Hub](https://create.roblox.com/docs/cloud/reference/features/luau-execution), Roblox Creator Hub (accessed 2026-10-01) - [Studio testing modes | Documentation - Roblox Creator Hub](https://create.roblox.com/docs/studio/testing-modes), Roblox Creator Hub (accessed 2026-10-01) ## Guardrails - Tests run only on a private test copy with test data. Live player data is never touched. - It never fixes bugs, and it never changes existing tests or requirements. - A real multiplayer playtest is run and signed off by a person, never by the agent. - Until your Ai1 team sets up the studio profile, the studio gives general Roblox help only and keeps no notes about the learner. ## Frequently asked questions ### Can automated game tests run without touching live players? Yes. The GS Playtester in Ai1 by MyZone AI runs automated tests on a private test copy of your Roblox game with test data, through Roblox's Open Cloud service and a stored key referred to by name only. Live player data is never touched or used: all game testing here runs on that private test copy. ### What makes a good bug report for a game? Exact steps to reproduce the problem, what should happen and what happened instead. The GS Playtester files a ticket like this for every failure, so builders see exactly what went wrong and can fix it the first time. It does not fix the bug itself. GS Producer sends the ticket to a builder. ### When does a Roblox game need a real multiplayer playtest? When a change affects how players join, leave or cope with lag. Those are hard to prove alone, so the GS Playtester flags the need. A person then runs the test with several players and signs it off before the change moves on. ### Does the GS Playtester work inside Roblox Studio, and when does it join? No. Automated tests run through Roblox's Open Cloud service. Anything that needs Studio is written as steps for the learner, who sends back screenshots and console output. It joins the studio from the second game. The learner reaches it through GS Producer, which sends builds here and brings the report back. ## About Ai1 Ai1 is the AI operations platform by MyZone AI, where each client runs on its own private server. The GS Playtester is not sold on its own: every Ai1 agent, including this one, is included on Developer Pro and every Fully Managed option, with no per-agent charge. Developer Core includes the development agents. Pricing: https://myzone.ai/pages/services/ai1-pricing. Security: https://myzone.ai/pages/security.