Query language models with voice commands. Helpful to automatically generate text, fix errors from dictation automatically, and generally speed up your Talon workflow.
-
See the list of prompts for all the prompts that can be used with the
modelcommand. -
See the examples file for gifs that show how to use the commands.
-
View the docs for more detailed usage and help
The OpenAI API that is used in this repo, through which you make queries to GPT 3.5 (the model used for ChatGPT), is not free. However it is extremely cheap and unless you are frequently processing large amounts of text, it will likely cost less than $1 per month. Most months I have spent less than $0.50
To add additional prompts, copy the Talon list for custom prompts to anywhere in your user directory and add your desired prompts. These prompts will automatically be added into the <user.modelPrompt> capture.
If you wish to change any configuration settings, copy the example configuration file into your user directory and modify settings that you want to change.
| Setting | Default | Notes |
|---|---|---|
| user.openai_model | "gpt-4o-mini" |
The model to use for the queries. NOTE: To access certain models you may need prior API use |
| user.model_temperature | 0.6 |
Higher temperatures will make the model more creative and less accurate |
| user.model_endpoint | "https://api.openai.com/v1/chat/completions" |
Any OpenAI compatible endpoint address can be used (Azure, local llamafiles, etc) |
| user.model_shell_default | "bash" |
The default shell for model shell commands |
| user.model_system_prompt | "You are an assistant helping an office worker to be more productive. Output just the response to the request and no additional content. Do not generate any markdown formatting such as backticks for programming languages unless it is explicitly requested." |
The meta-prompt for how to respond to all prompts |