AIToday
Large Language ModelsAI Coding AssistantsOpen-Source AIZenn AI/MLPublished: Oct 4, 2026, 10:00 JST

Minecraft AI bot plan hits 月数円 with Jev, Gemini, Mineflayer

Minecraft AI bot plan hits 月数円 with Jev, Gemini, Mineflayer

3 Key Points

  1. What happened

    The developer published fk2000/minecraft-ai-bot, which uses Jev to assemble instructions, Gemini to pick from allowed actions, and Mineflayer to hold the Minecraft Java Edition connection.

  2. Why it matters

    The build limits Gemini to choosing from three allowed actions — say, follow, and stop — so the AI cannot run arbitrary commands inside the game, which the author presents as reducing the risk of executing raw model output.

  3. What to watch

    The few-yen-a-month goal assumes an existing machine runs the bot and every service stays within its free tier, so the test is whether actual API usage and runtime stay under those limits. The author says the repository ships no bot implementation or Jev API definitions, so buildPrompt and similar parts must be replaced for the real Jev interface.

WHO IT HITSIndependent developers and hobbyists who want to run an AI-controlled Minecraft bot on a home machine are the main audience, since the design targets reusing existing hardware and free tiers rather than paying for new hosting.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, in your inbox every morning.

Context & Analysis

The design reflects a common split among hobbyist AI projects: one layer handles the instruction and conversation flow, another picks among allowed moves, and a third keeps a long-lived game connection. Here that means Jev formats the user's natural-language request together with the bot's role, current situation, and available actions; Gemini then selects one action and returns it as structured JSON rather than prose; Mineflayer executes the approved action.

The author is explicit that this is a design example rather than a working product. The repository includes no bot implementation or Jev API definitions, so buildPrompt and similar code must be adapted to Jev's actual interface. Safety choices are central: inputs are length- and character-limited, untrusted chat is not treated as commands, and the bot side re-validates whatever reaches it before calling the Mineflayer API. Failures such as network errors, empty responses, or broken JSON result in no action at all.

On cost, the few-yen-a-month figure is described as a target that only holds when an existing machine runs the bot and each service stays within its free tier. Renting new hardware can push the bill past that on its own, and the author notes that Gemini and Cloudflare free tiers, pricing, and available models change, so conditions should be checked before launch. Whether the cheap-operation claim holds for a given reader therefore hinges on usage volume, uptime, and which machine is already on hand.

FAQ
What does the bot actually let the AI do in Minecraft?
Gemini may only choose among a small pre-approved set of actions — say, follow, and stop — instead of generating free-form text. Anything outside that set, or any response that fails schema validation, is not executed.
Does the published repository contain a working bot and Jev API definitions?
No. The author states the repository does not include the bot implementation or Jev API definitions, and does not claim that a specific Jev SDK or API works as-is. The buildPrompt and similar parts must be swapped to match Jev's real interface.
Can I run the bot on Cloudflare Workers?
No. Mineflayer is a Node.js process that keeps a connection to the Minecraft server, so the bot cannot run continuously inside Cloudflare Workers. The author recommends separating the bot into its own Node.js environment.

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.

Questions and answers are published on this page.

Related Articles

Next articleBlackRock: stablecoins to power AI agent payments