Personal Telegram bot with a comprehensive AI assistant platform:

🤖 AI Model Support

• OpenAI Models: GPT-4, GPT-5, O3, O4-mini, DALL-E, Whisper, TTS
• Anthropic Models: Claude 3/3.5, Claude Haiku/Sonnet/Opus 4.x series
• Local Models via Ollama: Qwen, Gemma, open-source GPT models
• Replicate Models: Stable Diffusion, Flux Pro, Kandinsky, Llama, and 50+ others

🛠️ Advanced Features

• Function Calling & Tools: LLM models can call functions, generate images, search web
• Memory System: AI remembers user preferences and information across conversations
• Video Generation: Google Veo 3.1 models with context-aware audio
• Voice Processing: Speech-to-text with Whisper, multi-modal conversations
• Image Analysis: Upload and analyze images with AI
• Permissions System: Granular access control per user and feature
• 17 AI Personas: Specialized roles from IT Expert to Stand-up Comedian
• Video Downloads: Support for various video hosting platforms
• Meta-language Scheduler: Automated task execution with custom scripting language

🌐 Web App Builder

The bot can generate and host fully functional web applications directly in Telegram:

• Just describe what you want in plain language — the bot produces a complete self-contained HTML/JS/CSS app
• Apps open as Telegram Mini Apps inside the chat — no external browser needed
• Iterate with follow-up messages: "add a dark theme", "export to CSV", "show a bar chart"
• Built-in REST APIs available to apps running in the browser:
  • POST /api/llm — call any LLM on behalf of the current user
  • POST /api/replicate — run image/video generation models
  • GET /api/models — list models the user is allowed to use
• Apps can be shared with other users via a Telegram deep-link
• Manage created apps via 📱 My Web Apps in the main menu

Example prompts:

Create a dark-themed calculator
Make a habit tracker with weekly stats chart
Build a quiz app about world capitals with 10 questions
Create an AI text summariser that calls the bot's LLM API

Setup (Optional)

To enable the Web App Builder, configure the following in config.json:

• WEBAPP_BASE_URL — publicly accessible base URL where the server is reachable (e.g. https://mybot.example.com)
• WEBAPP_PORT — port for the built-in aiohttp server (default: 8180)

Here are a few simple steps to set it up

1. Clone the repository onto the server
2. Copy the config.json.example file to config.json
3. Register a Telegram bot
   • Send the command /newbot to the @BotFather bot
   • Enter the bot's name
   • Save the received token in config.json -> TELEGRAM_TOKEN
4. Add the Telegram user IDs of users who will have access to the bot to config.json -> ALLOWED_USER_IDS
   • Format: {"user_id": "Display Name", ...}
   • To get your Telegram user ID, message @userinfobot or @getmyid_bot on Telegram
   • Example: "ALLOWED_USER_IDS": {"123456789": "John Doe", "987654321": "Jane"}

If you need to use OpenAi(Optional):

5. Get the key from OpenAi:
   • Register at https://platform.openai.com/signup
   • Create a key at https://platform.openai.com/api-keys
   • Save the received key in config.json -> OPENAI_API_KEY

If you need to use Anthropic Claude models(Optional):

6. Get the key from Anthropic:
   • Register at https://console.anthropic.com/
   • Create a key at https://console.anthropic.com/account/keys
   • Save the received key in config.json -> ANTHROPIC_API_KEY

If you need to use replicate models(Optional):

7. Get the key from replicate:
   • Register at https://replicate.com/signin
   • Create a key at https://replicate.com/account/api-tokens
   • Save the received key in config.json -> REPLICATE_API_KEY

If you need to use local Ollama models(Optional):

8. Install and configure Ollama:
   • Download from https://ollama.ai/download
   • Pull desired models: ollama pull qwen3:32b
   • Set OLLAMA_HOST in config.json (default: http://localhost:11434)

Permissions Configuration(Optional):

9. Configure user permissions in config.json -> USER_PERMISSIONS:

   • Use user's Telegram ID as the key (must match ALLOWED_USER_IDS)
   • Set is_admin, can_use_tools, can_use_ollama_llm_models per user
   • The default key provides fallback permissions for users without specific settings
   • Control access to specific features and models
   • Example: "USER_PERMISSIONS": {"123456789": {"is_admin": true}, "default": {"is_admin": false}}

10. Install dependencies, and start the bot by running the command:

run.sh