s-nagaev / hiroshi

Just a GPT4Free-based Telegram bot providing 100% FREE GPT experience.

Hiroshi is a GPT4Free-based Telegram chatbot that offers 100% free access to interact with GPT-3.5, GPT-4, and Llama2 language models, inclusive of Bing, You, AI Chat, and more. Users have the freedom to select their preferred model or specific provider. Do note, the speed/stability may be slightly diminished when working with certain providers. Conversation context is fully retained when you switch between models and providers.

Note: This bot provides access to public free services. The quality and speed of such services can vary depending on various conditions and their current load. If you need a bot that uses the official OpenAI API and you have an API KEY, please check the following repository: https://github.com/s-nagaev/chibi.

Supported platforms

• linux/amd64
• linux/arm64 (Raspberry Pi 4+ is supported!)

Features

• Access to language models such as GPT-3.5, GPT-4, Llama2.
• Ability to choose a specific model or provider, with the fastest and most available provider automatically selected for the chosen model.
• Customizable storage time for conversation history.
• Preservation of conversation history even when changing providers or models.
• Pre-configured for quick setup, requiring only a Telegram bot token to get started.
• Cross-platform support (amd64, arm64).
• MIT License.

Can I try it?

Sure! @hiroshi_gpt_bot

System Requirements

The application is not resource-demanding at all. It works perfectly on the minimal Raspberry Pi 4 and the cheapest AWS EC2 Instance t4g.nano (2 arm64 cores, 512MB RAM), while being able to serve many people simultaneously.

Prerequisites

• Docker
• Docker Compose (optional)

Getting Started

Using Docker Run

1. Pull the Hiroshi Docker image:

   docker pull pysergio/hiroshi:latest

2. Run the Docker container with the necessary environment variables:

   docker run -d \
     -e TELEGRAM_BOT_TOKEN=<your_telegram_token> \
     -v <path_to_local_data_directory>:/app/data \
     --name hiroshi \
     pysergio/hiroshi:latest

   Replace <your_telegram_token> and <path_to_local_data_directory> with appropriate values.

Using Docker Compose

1. Create a docker-compose.yml file with the following contents:

      version: '3'
   
      services:
        hiroshi:
         restart: unless-stopped
         image: pysergio/hiroshi:latest
         environment:
           TELEGRAM_BOT_TOKEN: <your_telegram_token>
         volumes:
           - hiroshi_data:/app/data
      
      volumes:
        hiroshi_data:

   Replace <your_telegram_token> with appropriate values.

2. Run the Docker container:

   docker-compose up -d

Please, visit the examples directory of the current repository for more examples.

Configuration

You can configure Hiroshi using the following environment variables:

Variable	Description	Required	Default Value
TELEGRAM_BOT_TOKEN	Your Telegram bot token	Yes
ALLOW_BOTS	Allow other bots to interact with Hiroshi	No	false
ANSWER_DIRECT_MESSAGES_ONLY	If True the bot in group chats will respond only to messages, containing its name (see the BOT_NAME setting)	No	true
ASSISTANT_PROMPT	Initial assistant prompt for OpenAI Client	No	"You're helpful and friendly assistant. Your name is Hiroshi"
BOT_NAME	Name of the bot	No	"Hiroshi"
GROUP_ADMINS	Comma-separated list of usernames, i.e. "@YourName,@YourFriendName,@YourCatName", that should have exclusive permissions to set provider and clear dialog history in group chats	No
GROUPS_WHITELIST	Comma-separated list of whitelisted group IDs, i.e "-799999999,-788888888"	No
LOG_PROMPT_DATA	Log user's prompts and GPT answers for debugging purposes.	No	false
MAX_CONVERSATION_AGE_MINUTES	Maximum age of conversations (in minutes)	No	60
MAX_HISTORY_TOKENS	Maximum number of tokens in conversation history	No	1800
MESSAGE_FOR_DISALLOWED_USERS	Message to show disallowed users	No	"You're not allowed to interact with me, sorry. Contact my owner first, please."
PROXY	Proxy settings for your application	No
REDIS	Redis connection string, i.e. "redis://localhost"	No
REDIS_PASSWORD	Redis password (optional)	No
RETRIES	The number of retry requests to the provider in case of a failed response	No	2
SHOW_ABOUT	Just set it to false, if for some reason you want to hide the /about command	No	true
TIMEOUT	Timeout (in seconds) for processing requests	No	60
USERS_WHITELIST	Comma-separated list of whitelisted usernames, i.e. "@YourName,@YourFriendName,@YourCatName"	No
MONITORING_URL	Activates monitoring functionality and sends GET request to this url every MONITORING_FREQUENCY_CALL seconds.	No
MONITORING_FREQUENCY_CALL	If monitoring functionality is active sends GET request to MONITORING_URL every MONITORING_FREQUENCY_CALL seconds.	No	300
MONITORING_RETRY_CALLS	Logs error response only after MONITORING_RETRY_CALLS tries.	No	3
MONITORING_PROXY	Monitoring proxy url.	No

Please, visit the examples directory for the example of .env-file.

Versioning

We use SemVer for versioning. For the versions available, see the tags on this repository.

License

This project is licensed under the MIT License - see the LICENSE.md file for details.

hiroshi's Issues

[Feature Request] Append reply message to input

Description

When a user replies to other user message with @BotName , append the reply message to bot input.

Example:

User1 says:

eget nunc scelerisque viverra mauris in aliquam sem fringilla ut morbi tincidunt augue interdum velit

User2 replies to User1 text with: @BotName aliquet eget sit amet

Note: Lorem Ipsum used as example.

Current behavior

Bot only replies to User2 text without previous text replied message.

Proxy url

In what format should I add an url proxy?
And where i can find .env example file?

Empty Response/MissingAuthError/RateLimitError

I got all of this errors with 3 installation on 3 different PC with 3 different location including Stockholm and Germany.

Error after install in docker

Message in Telegram
"I'm sorry, but there seems to be a little hiccup with your request at the moment 😥 Would you mind trying again later? Don't worry, I'll be here to assist you whenever you're ready! 😼"

Log:
PermissionError: [Errno 13] Permission denied: '/app/data/########

got an error: [Errno 13] Permission denied: /app/data/'telegram_id'

Get error in docker run command
docker run \ -e TELEGRAM_BOT_TOKEN=token \ -v /home/gptbot:/app/data \ --name hiroshi \ pysergio/hiroshi:latest

Some Provider got error How to resolve?

Here is error
got an error: 403, message='Forbidden', url=URL('https://you.com/api/streamingSearch?userFiles=&selectedChatMode=default')

Added proxy still no luck

Installation complete. How do I proceed to run it?

[Feature Request] Environment variable for initial GPT provider

Description

Environment variable to set the initial GPT provider when (re)start the bot instead of default.

Example:
PROVIDER="ChatBase (GPT-3.5)"

CaptchaChallenge: Use other cookies or/and ip address

i got this error
CaptchaChallenge: Use other cookies or/and ip address

[Feature Request] Restrict privileged command to bot admins

Description

In a group or private, any user can use privileged commands (/reset, /provider)

Restrict privileged command to bot admin(s) only.

Example:
ADMINS="@User1,@User2"

UNHEALTH PORT CHECKING

Why does the port detection fail when configuring KOYEB, and the results show that it is unhealthy and cannot run. Which port do I need to EXPOSE? How to set it up?

Implementing an optional Integration with Monitoring Systems

Description:

The objective of this task is to enhance our chatbot application by providing an optional feature that allows it to connect with monitoring systems. This will enable the chatbot application to periodically report its operational status to the monitoring system by invoking a specific webhook.

Detailed Requirements:

1. Environment Variable Configuration:

   • Introduce two environment variables: MONITORING_URL and MONITORING_FREQUENCY_CALL.
   • The MONITORING_FREQUENCY_CALL variable should have a default value of 5 (the unit of time this refers to should be defined, e.g., minutes).

2. Operational Behavior:

   • Upon startup, if the MONITORING_URL variable is set, the chatbot application should begin making calls to the specified URL at intervals determined by the MONITORING_FREQUENCY_CALL.
   • These calls should commence immediately upon the application’s launch and continue at the specified frequency for as long as the application runs.

3. Logging:

   • Failed attempts to invoke the webhook should be logged, including the response code or type of error encountered. This will assist in troubleshooting and ensure transparency regarding the integration's functionality.
   • Successful calls need not be logged, to prevent cluttering the log files with routine operational entries.

4. Startup Notification:

   • At application startup, include a log entry that indicates the status of this monitoring feature. This should clearly state whether the feature is active (and if so, the configured values) or inactive. This initial log will serve as immediate feedback on the application’s configuration regarding its integration with external monitoring solutions.

Objective:

By accomplishing this task, we aim to improve the reliability and observability of our chatbot application. This feature will facilitate proactive maintenance and issue resolution by ensuring continuous communication between the chatbot and integrated monitoring systems.

[Feature Request] Whitelist topics

Description

Similar to GROUPS_WHITELIST but for group topics.
The bot can be added to a group, but restricted to interact on a specific topic(s) so don't spam other topics.

References:

• https://docs.python-telegram-bot.org/en/v20.7/telegram.forumtopic.html#telegram.ForumTopic
• https://core.telegram.org/bots/api#forumtopic
• https://telegram.org/blog/topics-in-groups-collectible-usernames

MissingRequirementsError

This is same problem? With blocked IP?
And nothing can do with that? Only change machine?
2024/04/25 14:12:59 stdout Chatgpt4Online: MissingRequirementsError: Install Webdriver packages | pip install -U g4f[webdri�[0m 2024/04/25 14:12:59 stdout Bing: RuntimeError: Empty response: Failed to create conversation 2024/04/25 14:12:59 stdout You: CloudflareError: Response 403: Cloudflare detected