Interact with Large Language Models (LLMs) via email.

• Utilize any OpenAI-compatible API with any available model
  • At the time of writing, OpenRouter offers free access to specific models with an OpenAI-compatible API
  • Ollama has an OpenAI-compatible API and allows for easy local deployment of LLMs
• Customize the behavior of the LLM with a system prompt
• Easily run in a Docker container
  • The default docker-compose.yml file uses restart: unless-stopped to ensure the container restarts after a reboot or if it crashes
• Check every n seconds
• No need for a local database - uses IMAP
• Use phidata for real-time information retrieval
  • Use Exa or DuckDuckGo for searching the internet
  • Get information from a URL when requested and understand it when requested

• Access to an OpenAI-compatible API (including locally hosted LLMs)
• An email account that supports IMAP and SMTP

• Python ^3.11
• Poetry (optional)

• Docker

This assumes you have the correct version of Python installed

1. pip install llmail
   a. You may need to use pip3 instead of pip
2. llmail

Cloning the repository is not required when installing from PyPi but is required when installing from source

1. Clone this repo with git clone https://github.com/slashtechno/llmail
2. cd into the cloned repository
3. Install and run with one of the following methods:

1. poetry install
2. poetry run -- llmail

1. Configure a .env file (see Configuration below)
   a. You can also edit the docker-compose.yml file directly but by default it loads the .env file
2. docker compose up -d

To configure the program, either use CLI flags (--help for more information) or environment variables (view .env.example for more information). It is recommended to just copy .env.example to .env and fill in the necessary information.

Once the program is running, you can send an email to the address you configured with whatever the subject is set to. The body of the email will be sent to the LLM, and the response will be sent back to you.
For example, if in .env you set SUBJECT_KEY=llmail, you would send an email with the subject llmail to the configured email address.

• If you used pip, run pip uninstall llmail
• If you used Poetry, just delete the virtual environment and then the cloned repository
• If you used Docker, run docker compose down in the cloned repository and then delete the cloned repository