A simple command line tool to generate text using any LLM (via litellm) based on ready made templated prompts.

heygpt stream

• To provide a simple command line tool to generate text using any LLM (OpenAI, Anthropic, Gemini, etc.) based on ready made templated prompts, in both cli as well as web-ui interface.

• CLI demo

• UI demo

• There is an optional dependency of fzf for interactive prompt selection. You can install it using your package manager.
• refer: fzf README for more info on steps to install fzf.

pip install heygptcli

heygpt --help

For debug logs use: export LOG_LEVEL=DEBUG or set LOG_LEVEL=DEBUG on windows.

You will need an API key from your LLM provider to use heygpt. Supported providers include OpenAI, Anthropic, Google (Gemini), and many more via litellm.

# gpt custom prompts (optional)
GPT_PROMPT_URL=<url-to-your-prompt-file>

# API Configuration (via litellm)
OPENAI_API_KEY=<your-api-key>  # Works with OpenAI and OpenAI-compatible endpoints
MODEL=gpt-4o  # optional model name (default: gpt-3.5-turbo)
              # Supports: gpt-4o, claude-3-5-sonnet, gemini-pro, etc.

In order to configure them you can use heygpt config command:

❯ heygpt config --help

 Usage: heygpt config [OPTIONS]

 Configure heygpt.

╭─ Options --------------------------------------------------------+
│ --prompt-file         TEXT  Prompt file path.                    |
│ --prompt-url          TEXT  Prompt file url.                     |
│ --api-key             TEXT  API key (supports any LLM provider). |
│ --api-endpoint        TEXT  API endpoint URL.                    |
│ --model               TEXT  LLM model name.                      |
│ --help                      Show this message and exit.          |
-------------------------------------------------------------------+

Default model name is gpt-3.5-turbo for this tool. You can change it to any model supported by litellm (e.g., gpt-4o, claude-3-5-sonnet-20241022, gemini-pro, etc.).

heygpt config --api-key <your-api-key>
heygpt config --model gpt-4o

Prompt YAML formate

# ~/path/to/prompts.yaml
- Title: Fix Grammar
  Command:
    - role: user
      content: |
        Review the provided text and correct any grammatical errors. Ensure that the text is clear, concise, and free of any spelling mistakes.

To use your saved prompts run:

heygpt config --prompt-file ~/path/to/prompts.yaml

Here, --prompt-url and --prompt-file is optional. If you want to use own custom prompts.

For providing a URL of yaml file containing your prompts.

# remote yaml file
heygpt config --prompt-url <url-to-your-prompt-file.yaml>

Note: This is the default yaml used for prompts: default-prompts.yaml, for using your own prompts, you need to follow the same format as in this file.

For your own prompts by providing a URL to a yaml file containing your prompts. You can also use local yaml file by providing a relative path to it.

# local yaml file
heygpt config --prompt-file ~/path/to/prompts.yaml

# ~/.config/heygpt/config.yaml
# You can manually add list of `available_models` in config file for easy access in streamlit UI.

- api_key: sk-proj-********
  model: gpt-4o
  available_models:
    - gpt-4o
    - chatgpt-4o-latest
    - gpt-4o-mini
    - gpt-3.5-turbo
    - claude-3-5-sonnet-20241022
    - gemini-pro
  prompt_file: /home/user/.config/heygpt/prompt.yaml

heygpt ask

• heygpt will ask you to choose a prompt from a list of available templates.

• After that, it will ask you to enter your query/task and will provide you with the result based on type of prompt you selected.

• You can specify a different model using the --model flag:

heygpt ask --model claude-3-5-sonnet-20241022
heygpt ask --model gemini-pro

• For asking queries without any prompt templates you can use --no-prompt flag.

heygpt ask --no-prompt

heygpt wisper ../path/to/audio.mp3

• You can provide standard output as well to heygpt ask

  echo "why sky is blue" | heygpt ask --no-prompt

  An other way to use it can be providing wisper audio 2 text, output to heygpt ask:

  heygpt wisper ../path/to/audio.mp3 | heygpt ask

This will start a streamlit server on localhost:

heygpt api

This will start a fastapi server on localhost: