🤖 Duke Bot

A self-hosted Telegram AI bot with Google Gemini & OpenAI-compatible APIs.
Deploy in 60 seconds with one click — no coding required.

---

✨ Features

Feature	Details
🧠 Multi-model AI	Google Gemini 2.0 Flash, 1.5 Pro, 1.5 Flash + any OpenAI-compatible provider
🔄 Model switching	/model — swap between AI models instantly via inline keyboard
🎭 System prompts	/prompt — Default, Professional, Creative presets or write your own
🖼️ Image understanding	Send photos — the bot analyzes and describes them
💬 Conversation history	Remembers context within each session
👥 Group chat support	Works in groups and supergroups; shared or per-user history
🔒 Access control	Restrict to specific users or groups to protect your API keys
🌐 Any OpenAI-compatible API	OpenRouter, Groq, Together AI, Mistral, self-hosted LLMs
⚡ One-click deploy	Vercel, Railway, or Render — no server required

---

🚀 Quick Deploy (Vercel)

3 steps — under 60 seconds:

1. Click Deploy with Vercel above → fill in your API keys when prompted
2. Visit your deployment URL → click Set Webhook on the setup page
3. Open your bot in Telegram → send /start

⚠️ Set ALLOWED_USERS to your Telegram ID during deploy to protect your API keys from abuse. Send /get_my_info to the bot to find your ID.

---

📋 Commands

Command	Description
/start	Welcome message and quick help
/new	Start a fresh conversation
/model	Switch AI model via inline keyboard
/prompt	Change system prompt (presets + custom)
/get_my_info	Get your Telegram user ID
/get_group_info	Get the group chat ID (groups only)
/help	Show full help message

Admin-only (requires ADMIN_ID + IS_DEBUG_MODE=1):

Command	Description
/get_allowed_users	List authorized users
/get_allowed_groups	List authorized groups
/get_api_keys	Show API key status
/list_models	List configured models

---

🤖 System Prompts

Preset	Description
🤖 Default	Friendly, concise AI assistant
💼 Professional	Expert consultant with structured answers
✨ Creative	Imaginative partner with vivid storytelling
✏️ Custom	Type your own system prompt

---

🌐 Supported AI Providers

Provider	Type	Base URL
Google Gemini	Gemini API	(built-in)
OpenAI	OpenAI-compatible	https://api.openai.com/v1
OpenRouter	OpenAI-compatible	https://openrouter.ai/api/v1
Groq	OpenAI-compatible	https://api.groq.com/openai/v1
Together AI	OpenAI-compatible	https://api.together.xyz/v1
Mistral	OpenAI-compatible	https://api.mistral.ai/v1
Self-hosted (Ollama, LM Studio, etc.)	OpenAI-compatible	Your server URL

---

⚙️ Environment Variables

Required

Variable	Description
BOT_TOKEN	Your Telegram bot token from @BotFather
GOOGLE_API_KEY	Google Gemini API key from Google AI Studio

Security (Strongly Recommended)

Variable	Description
ALLOWED_USERS	Comma-separated Telegram user IDs or @usernames. Set this to prevent strangers from draining your API quota.
AUCH_ENABLE	1 = only allowed users can chat (default), 0 = open to everyone

OpenAI-Compatible API (Optional)

Variable	Description	Default
OPENAI_API_KEY	API key for any OpenAI-compatible provider	—
OPENAI_BASE_URL	Provider base URL	https://api.openai.com/v1

Models (Optional)

Variable	Description	Default
DEFAULT_MODEL	Model used for new conversations	gemini-2.0-flash
GEMINI_MODELS	Comma-separated Gemini models	gemini-2.0-flash,gemini-1.5-pro,gemini-1.5-flash
OPENAI_MODELS	Comma-separated OpenAI-compat models	gpt-4o,gpt-4o-mini

Admin & Debug (Optional)

Variable	Description	Default
ADMIN_ID	Your Telegram user ID — enables admin commands	—
IS_DEBUG_MODE	1 = send debug logs to admin	0
GROUP_MODE	1 = shared group history, 2 = per-user history in groups	1

---

🆔 How to Find Your Telegram ID

Method 1 — Use this bot: Send /get_my_info to your bot after deploying — it replies with your user ID.

Method 2 — Use @userinfobot: Open it on Telegram and send any message — it replies instantly with your ID.

For group chat IDs: Add your bot to the group and send /get_group_info — the bot replies with the group ID (a negative number like -1001234567890). Add it to ALLOWED_GROUPS.

---

🔧 Other Deployment Options

Railway

1. Fork this repository
2. Create a new Railway project → Deploy from GitHub repo
3. Add environment variables in the Railway dashboard
4. Set start command: flask --app api.index run --host 0.0.0.0 --port $PORT
5. Set your webhook URL to the Railway domain

Render

1. Fork this repository
2. Create a new Web Service on Render
3. Set Build Command: pip install -r requirements.txt
4. Set Start Command: flask --app api.index run --host 0.0.0.0 --port $PORT
5. Add environment variables and set your webhook to the Render domain

---

🏗️ Architecture

Telegram  ──webhook──►  Flask on Vercel / Railway / Render
                               │
                    ┌──────────┴──────────┐
                    ▼                     ▼
             handle_message        handle_callback
                    │                     │
          ┌─────────┴──────┐        user_settings
          ▼                ▼        (model + prompt)
       text/photo     commands
          │
    ChatManager
    (per-session history)
          │
    ┌─────┴──────┐
    ▼            ▼
Gemini API   OpenAI-compatible API

Chat history is in-memory and resets on cold start (expected on serverless). Use /new to clear history manually.

---

📝 Register Commands with BotFather (Recommended)

Send this to @BotFather using /setcommands:

start - Start the bot
new - Start a fresh conversation
model - Switch AI model
prompt - Change system prompt
get_my_info - Get your Telegram ID
get_group_info - Get group ID
help - Show help

---

⭐ Support

If this project helped you, please star the repo — it helps others find it!

---

📄 License

MIT — free to use, modify, and deploy.