README.md

BufferChatMemory — Conversation Buffer

BufferChatMemory is the simplest memory backend for tinygent. It keeps an in-memory buffer of all messages exchanged in a conversation.

• Concept: every message (user or agent) is appended to an internal BaseChatHistory object.

• Available methods:

  • save_context(message): store a new message in the buffer.
  • load_variables(): expose the current buffer as a single string, ready to be injected into the next prompt.
  • clear(): drop the buffer entirely and start fresh.

• How it works in practice:

  1. Memory starts empty.
  2. Each call to save_context appends a new message.
  3. load_variables lets you pass the full conversation back into the model for continuity.
  4. When you no longer need past context, clear resets everything.

This is a simple, in-process memory — nothing is persisted to disk or external storage. It’s designed for examples, prototypes, or any agent that only needs short-term conversational context.