POSIX-compliant shell utility for working with local LLM + voice models
Requirements
Whisper CLI (e.g. sudo apt-get install whisper)
Ollama CLI (e.g. sudo apt-get install ollama)
TTS CLI (e.g. sudo apt-get install tts)
To use your desired models with Ollama and TTS, configure them and note their model names as you would use with ollama run xxxx or tts --model_name xxxx. Upon running RunTTS for the first time, go into the configuration menu and enter in the model names you are using.
Install
Clone the repo into ~/runtts.
chmod +x ~/runtts/runtts
~/runtts/runtts to start the program.
About
This tool is a mashup of the whisper, tts, and ollama CLIs to provide a local utility for interacting with AI models. RunTTS keeps track of a running context as you continue prompting it, and when needed, conversations can be saved for later prompting.
Due to the variability of situations where models can be run, RunTTS utilizes streaming responses to produce audio clips as soon as newline-delimited content is ready. Furthermore, it will handle markdown style triple-backtick blocks, setting them aside to not be read aloud, but can be viewed as received.
Suggested labels
{ "label-name": "AI-tools", "description": "Tools for working with AI models in a local environment", "confidence": 84.96 }
RunTTS
POSIX-compliant shell utility for working with local LLM + voice models
Requirements
sudo apt-get install whisper
)sudo apt-get install ollama
)sudo apt-get install tts
)To use your desired models with Ollama and TTS, configure them and note their model names as you would use with
ollama run xxxx
ortts --model_name xxxx
. Upon running RunTTS for the first time, go into the configuration menu and enter in the model names you are using.Install
~/runtts
.chmod +x ~/runtts/runtts
~/runtts/runtts
to start the program.About
This tool is a mashup of the whisper, tts, and ollama CLIs to provide a local utility for interacting with AI models. RunTTS keeps track of a running context as you continue prompting it, and when needed, conversations can be saved for later prompting.
Due to the variability of situations where models can be run, RunTTS utilizes streaming responses to produce audio clips as soon as newline-delimited content is ready. Furthermore, it will handle markdown style triple-backtick blocks, setting them aside to not be read aloud, but can be viewed as received.
Suggested labels
{ "label-name": "AI-tools", "description": "Tools for working with AI models in a local environment", "confidence": 84.96 }