You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We need a unified framework/protocol/practices for having different services implementable by different backends.
Such services include:
search: Google, DuckDuckGo, Exa, etc.
voice: OpenAI, ElevenLabs, Play.ht, Heygen, Wellsaid, Web Speech API
speech reco: OpenAI, opensource whisper, whisper in AI services, Web Speech API
image generation: OpenAI, Prodia (cheap stable diffusion), many many more...
code execution: E2B, Google Code Execution (impl in Big-AGI 2.0), Google Colab API, ...
We'd benefit from a strategy and a framework to allow for configuration, auto-detection, auto-selection of the best, user configuration and overrides.
What we need:
a prototype architecture which works for any implementation
a strategy for selection of a common feature interface, or a superset of features, across a category
a way for the user to configure each service (could be specific, but better if shared and stored uniformly)
a way for the user to override configurations on a x_project, x_chat, x_something_else, dynamic and stored
a way for the system to auto-select the best implementation, and update the selection if a better implementation becomes available (or if the user has altered the selection)
We need a unified framework/protocol/practices for having different services implementable by different backends.
Such services include:
We'd benefit from a strategy and a framework to allow for configuration, auto-detection, auto-selection of the best, user configuration and overrides.
What we need:
Originally posted by @enricoros in #614 (comment)
The text was updated successfully, but these errors were encountered: