Have Gemini handle a turn-based Android game
A developer set up Gemini to operate an Android phone and finish a turn in Fire Emblem Heroes without human taps.
- What they did
- They made an Android app in AI Studio so Gemini could act on the phone directly, including opening apps, tapping, swiping, and reading XML UI elements. Using that setup, Gemini 3.7 Flash examined the game state in Fire Emblem Heroes, saw it was the player’s turn, and guided the developer’s stronger characters through a battle round.
- How it went
- The agent completed the round. The developer said Gemini 3.6 had not managed this before, instead getting lost in menus and then stopping. They also noted that the actions were not very efficient and that even simple steps could still take a minute or two.
- Tips
- This kind of agent use is best for small, well-defined tasks, not polished play. Expect delays and some unreliable behavior, and avoid letting it do anything with money or other irreversible effects.
Try it yourself with Gemini
On my Android phone, open [mobile game] and use screen control to play through [one battle / one turn / one level]. Inspect the screen at each step, make reasonable tactical choices toward [goal], and avoid any paid actions or changes that can’t be undone. Before any purchase, send, delete, or submit action, stop and ask me to approve it. When the session ends, give me a short recap of the outcome and the key moves you made.
Fill in the [brackets]. Agents can make mistakes, so review anything before it is sent, paid or deleted.
Get 3 agent uses like this every morning
With the day's top AI news. Free, and one click to unsubscribe.
Related agent uses
Build a text game around AI narration
A developer used Grok to narrate a choice-based game and said it handled story text better than Gemini and GPT.
Guide a game agent through Goat Simulator 3
SIMA 2 in Goat Simulator 3 can be nudged by a person giving live directions.
Build an agent to play Pokémon Blue from RAM and screenshots
Joel Zhang wired Gemini 2.5 Pro into a Game Boy emulator so it could read game state and press buttons in Pokémon Blue.
Test screenshot-to-click scaling in a simple web game
A developer used Gemini with Playwright to confirm that click coordinates lined up with the browser view.
Use Gemini to generate a match-3 game from one prompt
A single Gemini prompt produced a working Candy Crush-style game and a share link.
Build a browser 4X game with LLM help
A developer used Grok and Gemini to assemble Hexwardia, a browser-based 4X game, in two weeks.