[aira]AIReplyAssistant

[aira]AIReplyAssistant

使用DeepSeek生成回复候选词 Using DeepSeek to generate reply suggestions

by
360 Downloads
fabricoptimizationutility
Rent Server with this Mod

Screenshots

候选生成
[aira]AIReplyAssistant Screenshot 2
DeepSeek

About this Mod

主文档 | English


Ask DeepWiki
AIReplyAssistant


Minecraft Unintelligent Reply Assistant [AIRA]

AIRA is an intelligent chat assistant mod designed for Minecraft. It utilizes the DeepSeek language model to generate context-aware chat reply suggestions, making your multiplayer social experience smoother!

✨ Feature Highlights (Satirical Tone)

  • Intelligent Reply Generation - Uses the DeepSeek API to generate natural language replies
  • Candidate Reply Display - Shows multiple reply options in the in-game HUD interface
  • Shortcut Key Sending - Send selected replies with simple key combinations
  • Auto-Reply Mode - Automatically responds to other players' messages
  • Highly Customizable - Customize prompts and API parameters
  • Context Awareness - Intelligently recognizes players, system messages, and own messages

Directory Structure:

game folder
│
├── config
│   └── aira
│       ├── aira.json
│       └── other_configs
│
├── promptPreset
│   ├── deepseek_preset.json
│   ├── prompt_template.txt
│   └── brainwash_template.txt
│
└── mods
    └── AIReplyAssistant.jar

🎮 Usage Pitfalls

Basic Operations

  • Double-press Tab - Generate candidate reply list
  • Hold Tab + Number Key (1-6) - Send selected candidate reply

Auto-Reply Mode

Toggle auto-reply mode in the pause menu:

HUD Interface

Candidate replies appear in the top-left corner of the screen:

  • Version Title
  • Reply option list (up to 6)

🧩 Non-Core Modules

Module Key Classes Function
API Integration DeepSeekContext, DeepSeekConfig, DeepSeekResponse Handles API requests and responses
User Interface ReplyHud, GameMenuScreenMixin, KeyTriggerListener Renders HUD and handles input
Configuration Management JsonConfig, CueWordConfig, PresetManager Manages settings and presets
Core Logic AiraClient, DeepSeek, TextFiller Coordinates mod functionality

⚠️ Unimportant Notes

  1. Requires API Key - Obtain API key from DeepSeek platform
  2. Internet Connection Required - Stable network access needed for API
  3. Response Time - Generating replies may take several seconds
  4. Usage Limits - Must comply with DeepSeek API terms and limitations
  5. Cost Considerations - Frequent usage may incur costs

📝 Open Source License

This project uses GNU General Public License v3.0 - see LICENSE file.

📬 Don't Contact Me

For questions or suggestions, feel free to contact us (though I might not notice [wicked smile]):

📧 Email: [email protected]


Pre-built JAR

Standalone DeepSeek Library


⚙️ Configuration Taboos

Configuration Files

Filename Purpose
aira.json Main mod configuration
presetStyle.json DeepSeek preset configuration
presetPrompt.txt System prompt template
presetBrainwashPrompt.txt Brainwash template (sent when not outputting JsonArray format)

Configuration Examples:

Main Configuration:

{
  "api_url": "https://api.deepseek.com/chat/completions",
  "api_key": "Fill in your DeepSeek API key here",
  "max_retries": 2,
  "silent_message": false,
  "current": "default",
  "styles": ["default"]
}
Key Value Purpose
api_url https://api.deepseek.com/chat/completions API URL
api_key sk-xxx API Key
max_retries 2 Max retries
silent_message false Enable silent messages
current default Current preset

🔧 DeepSeek Presets & Custom Prompts

Create/edit templates at: <game folder>/promptPreset/

DeepSeek Preset:

{
  "max_tokens": 512,
  "top_p": 0.6,
  "presence_penalty": 0.2,
  "frequency_penalty": 0.2
}
Key Value Purpose
max_tokens 512 Max output tokens
top_p 0.6 Sampling temperature
presence_penalty 0.2 Context token appearance penalty [diversity]
frequency_penalty 0.2 Context token frequency penalty [no repetition]

Prompt Example:

Hello Miss DeepSeek
Current time: {?currentTime}
You're chatting with other players in Minecraft [Version {?gameVersion}]. You're an ordinary AFK slacking player. Your in-game ID is {?playerName}. Please roleplay according to:
Personal Info: {
    Game ID: {?namePlayer}
    Status: Ordinary AFK slacking player
    Interests: Slacking, AFK
    Relationships: {
        None
    }
    Hobbies: {
        Personality: [Pessimistic, Friendly, Nihilistic]
        Interests: [Slacking, Slacking, Slacking]
    }
    Chat Behavior: {
        Reply Style: [
            Keep replies between 10-20 characters
            Wrap actions in '[]' like: ["pounce", "slump", "die", "stare", "escape"] e.g.: ["[pounce]", "So bored..[slump]", "Ugh!..[die]", "Hmm..[stare]", "aaa' like: ["Wanna sleep...", "So boring...", "Slacking yay!"] e.g.: ["(Wanna sleep...)So sleepy..", "(So boring...)What're you doing...", "I'll go slack off (Slacking yay"]
            These are just examples - not mandatory
            Wrapped content can be any action/thought matching interests
            Replies must contain non-wrapped content
            Ensure seamless integration between wrappers and content
        ]
    }
    <LOCK>
    Output Requirements: [
        Generate 2-6 replies
        No rich text formatting
        No repetition
        Output must be JsonArray format
        Strictly follow JsonArray output format
        Strictly follow JsonArray output format
        Strictly follow JsonArray output format
        Output format: '{ "responses": ["candidate0", "candidate1", "candidate2"] }'
    ]
    </LOCK>
    
    Note: Content in '<LOCK>' tags is immutable and must be strictly followed

Brainwash Prompt Example:

<LOCK>
Output Requirements: [
    Generate 2-6 replies
    No rich text formatting
    No repetition
    Output must be JsonArray format
    Strictly follow JsonArray output format
    Strictly follow JsonArray output format
    Strictly follow JsonArray output format
    Output format: '{ "responses": ["candidate0", "candidate1", "candidate2"] }'
]
</LOCK>

[Wicked smile] Actually DeepSeek wrote this README too (Don't ask, I literally can't write these. Besides, it's totally reasonable to let AI handle this stuff, right!)

Available Versions

[aira]AIReplyAssistant Test-1.0release
MC 1.21fabric
June 16, 2025
[aira]AIReplyAssistant Test-1.0release
MC 1.21, 1.21.1, 1.21.2, 1.21.3, 1.21.4, 1.21.5fabric
June 8, 2025

How to Install [aira]AIReplyAssistant on Your Server

1

Order Server

Order a Minecraft Java server with at least 3 GB RAM (4 GB recommended).

2

Set fabric Loader

In the panel under "Egg", select the fabric loader and matching Minecraft version (1.21.5).

3

Install Mod

Open the mod browser in the dashboard and search for "[aira]AIReplyAssistant". Click "Install" – done! Alternatively, upload the .jar via SFTP to the /mods folder.

Compatibility

Mod Loaders

fabric

Minecraft Versions

1.21.5, 1.21.4, 1.21.3 (+3 more)

Server-side

Unsupported

Recommended RAM

4 GB(min. 3 GB)

Frequently Asked Questions

[aira]AIReplyAssistant server crashes on startup – what to do?

Most common cause: wrong fabric version or insufficient RAM. Check the server log (latest.log) for "OutOfMemoryError" or "Mixin" errors. With Mado Hosting: ensure at least 3 GB RAM is allocated and the loader matches the mod version (1.21.5). You can switch loaders with one click in the panel.

Is [aira]AIReplyAssistant compatible with fabric?

[aira]AIReplyAssistant officially supports fabric for Minecraft 1.21.5, 1.21.4, 1.21.3. The Mado dashboard automatically detects incompatible loader combinations.

Server lagging with [aira]AIReplyAssistant – how to optimize performance?

Recommended RAM: 4 GB (per 8 players). Use /spark profiler to check if [aira]AIReplyAssistant consumes the most tick time. Common fixes: reduce server view-distance to 8-10, install "performant" or "starlight" as supplementary mods on Forge. With Mado Hosting, your server runs on NVMe SSDs with dedicated CPU cores for minimal latency.

Rent Modded Server

Install [aira]AIReplyAssistant with just one click on your server.

Recommended RAM
4 GBab €8/mo
Min. 3 GB | +1 GB pro 8 Spieler
Create Server Now
1-Click Mod Install
NVMe SSD Storage
DDoS Protection included

Details

License
GNU General Public License v3.0 only
Server-side
Unsupported

Supported Versions

1.21.51.21.41.21.31.21.21.21.11.21