iMessage Chat Bot is a MacOS application that plugs the power of modern LLM's into iMessage.Use it as a funny tool to annoy friends, or something useful such as a personalized messaging system to inform your loved ones you are not available to chat. The chat bot uses your number to send the automated responses, simplifying the experience of automating and interacting with texts. You can use it a python script or install the app from releases.
The distributable app is a MacOS menubar app that will let you start and stop the chatbot running in the background.
The chat bot utilizes a json file to help configure the settings and behavior when running the program. An example format for the
config.json can be found here. Below is also a rundown of each of the sections and explanation
of their purpose.
| Name | Description |
|---|---|
| max_tokens | The number of tokens that an LLM can process in a single interaction. |
| delay_between_loops | The amount of time to wait until pinging the local chat database for updates. |
| personal_info | An object that contains information about you. This is provided to Chat GPT if context is set to true. |
| default_single_chat | a single chat object that is the default response settings for the auto-messenger. This is an optional section. If not provided, no default messages will be sent. |
| single_chats | A list of single chat objects that override the default_single_chat settings. Unique identifier is the phone_number |
| group_chats | A list of group chat objects that describe the settings for the auto-messenger. This is an optional section. If not provided, no responses will be made to group chats. Keys are the recipients name. |
| Name | Description |
|---|---|
| name | The name of the recipient. |
| phone_number | The phone number of the recipient. |
| about | A brief description of the purpose of the chat bot. This is not used as context for Chat GPT. |
| prompt | The prompt to send as a request to Chat GPT. |
| context | true to provide additional chat details to Chat GPT request. |
| only_reply_focus_mode | true to only reply in focus mode. |
| number_of_previous_messages | if context is true, sends the n previous messages as context to create a reply. This is constrained by the maximum number of tokens allowed. |
| emoji_pasta | Creates emoji-fied responses from the auto-messenger. |
| bot_trigger_command | The trigger command to enable the bot in a chat. If an empty string, replies will happen with all incoming texts. |
| Name | Description |
|---|---|
| recipients | A list of recipients that are apart of the group chat. Recipients' name and phone number must match for every individual in the chat. |
- MacOS computer
- iPhone with message forwarding enabled
- An iCloud account and iMessage setup
- Python3 (developed on 3.11)
- ChatGPT account with API Key
This has to be ran on mac with iCloud account and iMessage setup and thus will appear to be sent from you rather than an external number. Also I believe that groupchat has to be already created. I initially tried to use twilio to power the texts but it looks like as of right now (April 2023) they have deprecated group messaging.
coming soon...
- Create virtual environment with
python -m venv .venv - Enable your virtual env with
source .venv/bin/activate - Install python packages with
pip install -r requirements.txt - Install the py-imessage-shortcuts package. You'll need to set up the shortcut by double clicking the
send-imessage.shortcutand then selecting "Add shortcut" (the shortcut is from that repo - I did not create it). - Create a Chat GPT API Key here and add it to your config file. Alternatively, you can set the global variable in the command line:
export OPENAI_API_KEY=<your-api-key> - In order to read iMessage data, the application or terminal will need full disk access. iMessage data exists in
~/Library/Messageswhich is protected. Go toSettings -> Privacy & Security -> Full Disk Accessand select the application that should get access. - Ensure iMessage text forwarding is enabled on your iPhone.
- Run the bot and pass your config with:
python src --config configs/example-config.json
- Install pyinstaller (in your virtual environment):
pip install pyinstaller - Install (create-dmg)[https://github.com/create-dmg/create-dmg] onto you computer, easiest way is through the brew package manager (which you'll have to install seperately if you don't have it)
brew install create-dmg - Run the
make_dmg.shscript. you may have to make it executablechmod +x make_dmg.sh, then./make_dmg.sh. This script runs both py installer and create-dmg to built the distributable .dmg file. - Find the .dmg inside the
dist/folder.