Skip to main content

AI Features

grafychat is designed to take advantage of the full range of AI Chat capabilities and currently supports the following AI features:

Chat​

Any activated model can be used as a chat model.
Some models (like GPT-4) support custom instructions.
Prompt for each chat is located in Node's Request.
The Selected Model will generate response in Node's Response.

Summarizer​

Summarizer is a convenient method to summarize data into Node's Label.
Any activated model can be used as a summarizer model. The prompt, model and temperature for summarizer can be configured per context in Context Settings.

Voice Input​

  • We currently support OpenAI's whisper-1 to transcribe voice to text.
  • Please allow microphone access in your browser. Note: this feature is best utilized in Chrome or Firefox.

Text to Speech​

  • We currently support OpenAI's tts-1 model to convert text to speech.
  • You can choose between different voices.
  • The generated audio can be downloaded as MP3.

Image Generation (Experimental)​

  • We currently support OpenAI's DALL-E-3 to generate images.
  • Enter your prompt in Node's Request.
  • Output image will be displayed in Node's Response with a download link and revised prompt generated by DALL-E-3 under the hood.

Note: grafychat does not store images and generated image will expire in less than an hour. To be able to keep the image you can upload in to any cloud storage and update the link in the markdown.

Text Embedding (In Development)​

  • We are working on something special for this.