0xAI
ModMinecraft mod that adds AI chat via Ollama. Type `@ai` in chat to ask questions.
Description
oxai
oxai is a Fabric client mod that adds an Ollama-powered chat companion to Minecraft. Type @ai in chat and the mod sends the prompt to your local Ollama server, then replies in chat.
Support
- Mod version:
0.1.5 - Minecraft:
26.1.x - Fabric Loader:
0.17.3+ - Fabric API: required
- Java:
25+ - Mod Menu: optional, supported
Mod Menu adds an in-game config screen for the same settings stored in config/oxai.json.
Install
- Install Fabric Loader and Fabric API for Minecraft
26.1.x. - Put the 0xAI JAR in your
modsfolder. - Optional: install Mod Menu to configure 0xAI in-game.
- Start Ollama:
ollama pull llama3.2
ollama serve
- Launch Minecraft.
Usage
In chat:
@ai how do i find diamonds
@ai what's the best food
@ai how do i make a nether portal
The default trigger is @ai. You can change it in Mod Menu or in config/oxai.json.
Configuration
With Mod Menu installed:
- Open Mods.
- Select 0xAI.
- Open the config screen.
- Change settings and press Done.
Without Mod Menu, edit config/oxai.json:
{
"enabled": true,
"trigger": "@ai",
"ollamaUrl": "http://localhost:11434",
"model": "llama3.2:3b",
"systemPrompt": "You are a helpful AI assistant in a Minecraft server. Keep responses concise and friendly.",
"maxTokens": 150,
"temperature": 0.7,
"responsePrefix": "[AI] ",
"maxMessageLength": 256,
"showErrors": true,
"requestTimeout": 60000,
"contextMessages": 5
}
| Option | What it does |
|---|---|
enabled |
Turns AI replies on or off |
trigger |
Chat text that triggers the AI |
ollamaUrl |
Ollama server address |
model |
Ollama model name |
systemPrompt |
System message sent with each request |
maxTokens |
Maximum generated response length |
temperature |
Response randomness from 0.0 to 1.0 |
responsePrefix |
Prefix added before AI chat messages |
maxMessageLength |
Maximum characters per chat line |
showErrors |
Shows request errors in chat |
requestTimeout |
Ollama request timeout in milliseconds |
contextMessages |
Number of recent chat turns to keep as context |
Models
0xAI works with any Ollama chat model. Common choices:
llama3.2:3bllama3.2:1bmistralphi3
Use smaller models for faster replies or lower-end hardware.
Build
Install JDK 25+, then build with Gradle:
./gradlew build
On Windows PowerShell, if the Unix wrapper script is not executable, run:
java -classpath gradle\wrapper\gradle-wrapper.jar org.gradle.wrapper.GradleWrapperMain build
The built JAR is written to build/libs/.
Troubleshooting
Build says release version 25 not supported
Gradle is using an older JDK. Install JDK 25 and make sure JAVA_HOME or your IDE Gradle JVM points to it.
The mod does not respond
- Make sure 0xAI is enabled in Mod Menu or
config/oxai.json. - Check that Ollama is running with
ollama serve. - Check that the configured model is installed with
ollama list. - Verify that your message contains the configured trigger.
Replies are slow
- Use a smaller model such as
llama3.2:1b. - Lower
maxTokens. - Lower
contextMessages.
Connection refused
- Check
ollamaUrl. - Make sure Ollama is running on the configured host and port.
Compatibility
Mod Loaders
Game Versions
Recommended Gear
AffiliateHardware and extras that pair well with this mod.
As an Amazon Associate I earn from qualifying purchases.
Similar Mods
Coffee Variants
Coffee Variants is a mod that adds a variety of coffee to minecraft as well as a barista villager and a coffee structure...
bewitched.rar
A horror mod, inspired by other, well-known ARGs, including horrendous entities. Your world is alive. Run.
Bewitchment REI
Adds REI support to Bewitchment for 1.20
Ant Man - Palladium/FSang
PymTech Remastered PymTech Remastered is a fan-made Minecraft mod inspired by the classic PymTech, bringing Ant-Man-styl...
Twilight Eye - End Remastered Add-on
This is an Add-on for End Remastered Mod Which Adds a new Eye in Twilight Forest Dimension Dropped by Twilight lich
Universal Circuits
Standalone Universal Circuits for GregTech, inspired by Nomifactory & Monifactory.