Discord Gemini Bot (Advanced, Anti‑Ban, Smart Conversations)

A Discord chat assistant with:

• Auto dependency install on first run
• Hot‑reloadable config.json
• Multiple Gemini API keys (from .env and/or config), automatic rotation with per‑key cooldown/backoff on 429
• Human‑like behavior, anti‑ban controls, per‑user cooldowns, and channel slowmode
• Smart conversations: waiting windows, alternating priority, 10 user cap
• Owner text commands and a live terminal dashboard with recent events

Note: For educational purposes. Please follow Discord ToS and local laws.

1) Requirements

• Python 3.10+
• Discord token
• One or more Gemini API keys (free or billing‑enabled)

2) Quick Start

Windows (easiest):

# A) Double‑click
run_bot.bat

# B) Or from terminal
python run_bot.py

Cross‑platform direct run:

python bot.py

This auto‑installs missing packages and starts the bot.

3) Environment (.env)

Create a .env file in the project root:

DISCORD_TOKEN=your_discord_token
GEMINI_API_KEY=your_primary_gemini_key
BOT_OWNER_ID=your_discord_user_id

# Optional: multiple keys (any of these formats work)
GEMINI_API_KEYS=key2,key3,key4
# or
GEMINI_API_KEYS=key2;key3;key4
# or multiline
GEMINI_API_KEYS=key2
key3
key4

• The bot reads keys from both .env (GEMINI_API_KEY + GEMINI_API_KEYS) and config.json.gemini_api_keys. Lists are combined and de‑duplicated.

4) Configuration (config.json)

config.json is hot‑reloaded every loop (no restart needed):

{
  "default_mode": "casual",
  "cooldown_seconds": { "min": 60, "max": 120 },
  "channels": {
    "default_slowmode": 5,
    "overrides": {
      "123456789012345678": 8
    }
  },
  "response_patterns": ["casual", "professional", "funny", "helpful", "quiet"],
  "webhook_url": "",
  "owner_id": "123456789012345678",
  "gemini_api_keys": ["key5", "key6"]
}

• default_mode: quiet/casual/professional/funny/helpful.
• cooldown_seconds: per‑user cooldown window (randomized between min–max).
• channels.default_slowmode: global refresh delay (seconds).
• channels.overrides: per‑channel slowmode override.
• gemini_api_keys: optional extra keys; combined with .env keys.

5) Running

On start the bot:

1. Installs missing requirements
2. Validates tokens/connectivity (Gemini 429 shows as a warning; bot still runs)
3. Prompts for channel IDs (comma‑separated numeric IDs)
4. Prompts for slowmode (or uses config/default)
5. Shows a live dashboard with stats and recent events

6) Key Features (current behavior)

• Multiple Gemini keys & backoff:
  • Reads keys from .env and config.json, deduped and rotated automatically
  • On 429/quota, marks current key on cooldown using retry_delay seconds (if provided) and switches to the next available key
• No greeting starts:
  • Leading “hey/hi/hello …” are stripped from AI replies
• Emoji placement:
  • Never prefix; prefers inline after a relevant keyword; otherwise suffix at the end
• Anti‑ban controls:
  • Per‑user cooldown (configurable, default 60–120s randomized)
  • Channel slowmode (configurable and per‑channel overrides)
  • Random skip (20%) to avoid over‑activity
  • Breaks, channel rotation, quiet mode pattern support
• Conversations:
  • Active session window: 5 minutes per user
  • While waiting for a conversation reply, the bot shows a live countdown and replies to new users without blocking
  • If the conversation user replies, the bot prioritizes that message immediately on the next poll
  • After slowmode, bot checks conversation user first; if no new message from them, replies to new users
  • Conversation cap: max 10 users; expired (>5 minutes inactive) or oldest sessions are removed to admit new users

7) Owner Commands (in‑channel text)

Only owner_id/BOT_OWNER_ID can use these:

!pause            # pause replies
!resume           # resume replies
!mode <name>      # quiet / casual / professional / funny / helpful
!slowmode <sec>   # set channel slowmode seconds
!status           # print dashboard to console
!rotate           # rotate to next channel (if configured)

8) Dashboard

• Header line shows: channel, mode, slowmode, next refresh ETA
• Stats: processed, responses, errors, response rate, active conversations
• Recent events: replies, mode/slowmode changes, rotations, errors, rate‑limits
• Shows a per‑second countdown when waiting for a conversation reply

9) Tips

• To reduce 429s, add multiple keys in .env/config.json, or use a billing‑enabled key
• Keep slowmode and per‑user cooldown conservative to mimic human pace
• Use quiet mode during busy hours

10) Troubleshooting

• 429/quota warnings: normal for free‑tier; bot rotates to next key automatically and cools down the throttled key
• Unauthorized (401) from Discord: check DISCORD_TOKEN
• Module not found: run python bot.py once (auto‑installs) or pip install -r requirements.txt
• Unicode/emoji on Windows: use run_bot.py or Windows Terminal

11) Project Files

• bot.py: main app (AI, logic, dashboard)
• config.json: hot‑reloaded settings
• .env: tokens and optional GEMINI_API_KEYS
• run_bot.py / run_bot.bat: one‑click launchers
• requirements*.txt: dependency lists
• install_requirements.py: optional helper (not required)

Enjoy building responsibly! 🎯