Skip to main content

Indie game storeFree gamesFun gamesHorror games
Game developmentAssetsComics
SalesBundles
Jobs
TagsGame Engines
A jam submission

Tap Turn BattlerView game page

Party dungeon crawler. Grid. Turns. Free to floor 50 in the browser.
Submitted by tapturnbattler (@tapturnbattler) — 4 days, 23 hours before the deadline
Add to collection

Play Browser Version

Tap Turn Battler's itch.io page

Results

CriteriaRankScore*Raw Score
Overall Enjoyment & Value#43.6673.667
AI Direction & Mastery#43.6673.667
Innovation & Experimentation#43.5003.500
Production Velocity & Scope#53.5003.500

Ranked from 6 ratings. Score is adjusted from raw score by the median number of ratings per game in the jam.

The Biggest Challenge
The hardest part was not generating assets — it was directing AI so the pieces still formed one playable game. Meshy gave us heroes, foes, hub buildings, and dungeon props fast, but getting them into a Three.js HTML5 build meant careful export (skins, clips, upright orientation) without breaking combat. On the code side, Cursor helped ship a full delve loop — hub, procedural floors, AP combat, fatigue, shop, saves — but the breakthrough was learning to steer it with tight constraints: fix the aim prompt, don’t rewrite combat; heal must raise HP, don’t invent new systems. The win was treating AI as a production partner under human direction, not as an autopilot.

AI Tools Used
Cursor (agent / Composer) for design, TypeScript/Three.js gameplay, debugging, and itch builds; Meshy for rigged characters, environment props, and dungeon kit meshes; occasional LLM chat for jam copy and design review.

Areas Where AI Was Used
Entire game is created with AI

Leave a comment

Log in with itch.io to leave a comment.

Comments

Submitted(+1)

I think this is pretty dang impressive!!!!! While controls are wonky and the audio is subpar, the mechanics of the game are really complex for what I've seen anyone do with ai. This is genuinely wonderful work. 

Developer

Thanks! I am hoping the "tapping" feels better after the next update :)

Submitted

1. Overall Enjoyment & Value: 

As someone who heavily prefers deep, text-driven narrative experiences over mindless action, Tap Turn Battler gave me mixed feelings. It’s a grid-based party dungeon crawler with heavy strategic turns, which scratches that methodical itch, but it lacks the rich storytelling and prose depth I look for. It is functional, playable, and has a clear tactical loop, but the overall engagement flatlines a bit due to repetitive mechanics and sparse narrative flavor.

2. AI Direction & Mastery: 

The game feels like standard AI utility usage. While the pixel art assets and UI elements function well together, they carry that unmistakable "off-the-shelf" procedural look without a deeply unique artistic voice shining through. It works, but it lacks the sophisticated prompt architecture or curated style that makes an AI-driven project truly stand out.

3. Production Velocity & Scope: 

For a jam scope, the utilization of generative tools clearly allowed the developer to push out a massive amount of content—stretching up to 50 floors in the browser version. Features like party management, custom UI layouts, and grid systems would have eaten up massive amounts of solo time without AI pipeline acceleration.

4. Innovation & Experimentation: 

Conceptually, it plays it relatively safe within established retro dungeon crawler tropes. While leveraging AI to rapidly iterate a deep tactical grid system is a neat feat, the core gameplay ideas don’t break entirely new ground compared to traditional entries in the genre.

Pipeline & Technical Evaluation:

 * Architecture & Pipeline: Signs point to automated asset workflows and structured JSON/data schemas to handle the sprawling floor counts and grid mechanics, though deep RAG integration for local lore is largely absent since text and story take a backseat to raw dungeon crawling.

 * Narrative & World-Building: Highly minimal. Dynamic branching and deep lore auditing take a back seat here to mechanical turn loops and combat pacing.

 * QA & Simulation: The game has seen steady iteration on UI polish, combat feel, and bug fixes across multiple devlogs, though rough edges like healing registration or door triggers still occasionally require manual player workaround.

Developer

Thank you for your post. I urge you to give my game another try when in my next update after this jam completes. I am hopeful the next update will make the game feel much better!