🦙 Ollama Telegram Bot

Chat with your LLM, using Telegram bot!
Feel free to contribute!

Features

Here's features that you get out of the box:

• Fully dockerized bot
• Response streaming without ratelimit with SentenceBySentence method
• Mention [@] bot in group to receive answer

Roadmap

• Docker config & automated tags by StanleyOneG, ShrirajHegde
• History and /reset by ShrirajHegde
• Add more API-related functions [System Prompt Editor, Ollama Version fetcher, etc.]
• Redis DB integration
• Update bot UI

Prerequisites

• Telegram-Bot Token

Installation (Non-Docker)

• Install latest Python

• Clone Repository

  git clone https://github.com/ruecat/ollama-telegram
  

• Install requirements from requirements.txt

  pip install -r requirements.txt
  

• Enter all values in .env.example

• Rename .env.example -> .env

• Launch bot

  python3 run.py
  

Installation (Docker Image)

The official image is available at dockerhub: ruecat/ollama-telegram

• Download .env.example file, rename it to .env and populate the variables.

• Create docker-compose.yml (optionally: uncomment GPU part of the file to enable Nvidia GPU)

  version: '3.8'
  services:
    ollama-telegram:
      image: ruecat/ollama-telegram
      container_name: ollama-telegram
      restart: on-failure
      env_file:
        - ./.env
    
    ollama-server:
      image: ollama/ollama:latest
      container_name: ollama-server
      volumes:
        - ./ollama:/root/.ollama
      
      # Uncomment to enable NVIDIA GPU
      # Otherwise runs on CPU only:
  
      # deploy:
      #   resources:
      #     reservations:
      #       devices:
      #         - driver: nvidia
      #           count: all
      #           capabilities: [gpu]
  
      restart: always
      ports:
        - '11434:11434'

• Start the containers

  docker compose up -d

Installation (Build your own Docker image)

• Clone Repository

  git clone https://github.com/ruecat/ollama-telegram
  

• Enter all values in .env.example

• Rename .env.example -> .env

• Run ONE of the following docker compose commands to start:

  1. To run ollama in docker container (optionally: uncomment GPU part of docker-compose.yml file to enable Nvidia GPU)

     docker compose up --build -d
     

  2. To run ollama from locally installed instance (mainly for MacOS, since docker image doesn't support Apple GPU acceleration yet):

     docker compose up --build -d ollama-tg
     

Environment Configuration

Parameter	Description	Required?	Default Value	Example
TOKEN	Your Telegram bot token. [How to get token?]	Yes	yourtoken	MTA0M****.GY5L5F.****g*****5k
ADMIN_IDS	Telegram user IDs of admins. These can change model and control the bot.	Yes		1234567890 OR 1234567890,0987654321, etc.
USER_IDS	Telegram user IDs of regular users. These only can chat with the bot.	Yes		1234567890 OR 1234567890,0987654321, etc.
INITMODEL	Default LLM	No	llama2	mistral:latest mistral:7b-instruct
OLLAMA_BASE_URL	Your OllamaAPI URL	No		localhost host.docker.internal
OLLAMA_PORT	Your OllamaAPI port	No	11434
TIMEOUT	The timeout in seconds for generating responses	No	3000
ALLOW_ALL_USERS_IN_GROUPS	Allows all users in group chats interact with bot without adding them to USER_IDS list	No	0

Credits

• Ollama

Libraries used

• Aiogram 3.x