Generation Stage
You may want to show your users something unique during each stage of the generation process, rather than a static 'loading' message for the entire duration.
To enable this, each thread has a generationStage
property that can be used to track the progress of the generation process while processing a message.
You can expose the generation stage ( or a value you derive from it ) to your users so they know what their AI assistant is doing.
Stages
A thread's generation stage is an enum with the following possible values:
Status Message
Instead of determining a custom message for each stage, you can use the statusMessage
property, which is a context-aware string that describes the current stage of the generation process, as generated by Tambo.
For example, if the generation was triggered by a user message asking about the weather in Tokyo, during the CHOOSING_COMPONENT
stage the statusMessage
might be something like "Looking for a component to help with your request about the weather in Tokyo"
.
Use this along with the isProcessing
helper property which returns true
if the generationStage
is CHOOSING_COMPONENT
, FETCHING_CONTEXT
, HYDRATING_COMPONENT
, or STREAMING_RESPONSE
.