Chat continuity support

[EncHawk]

Includes both client and backend interfaces

1. Server needs to be able to handle the new context from the chat

•  being able to take the context for every response. Can be done easily with the context table already
  

•  So for each response the model retrieves all context from the chat (sent from frontend; obtained from supabase)
  

2. Every chat node can be able to continue multimodal chat. (prolly doesnt need mutlimodal in a single chat, no point in it)

•  after one response, using the contxt from that chat, next conversation can be continued.
  

3. (optional) Merge chats, where the context from one chat can be passed on to the next ones.

•   prolly just a game of validation, that returns the context from a single chat node.
  

[EncHawk]

All we need to do is to be able to store context based on user && the node. Maybe two tables one for user chat, and one for text in a node can be used for this.

all that context or selective context from that in embeddings can be passed into the LLM somehow.