A simple canvas for language and audio models.

The article discusses the process of integrating third-party services to a user's canvas, emphasizing that all keys are stored locally and all calls to third parties occur directly between the user's browser and the service. It also mentions that during the beta phase, users won't need to enter any keys as these integrations will be provided.

The article further explains that users will be given the ability to publish and share their workflows. An example is provided where the five most recent items from an RSS feed are taken, a readable summary is created using `mixtral-8x7b-instruct`, and the output is then sent to an ElevenLabs tts model.

Key takeaways

Connect third party services to start adding blocks to your canvas, with keys kept in local storage.
All calls to third parties happen between your browser and the third party service.
In the beta, no keys are required as integrations are provided.
There will be a way to publish your workflows and share them with others.

A simple canvas for language and audio models.

Key takeaways

Discussion (0)