Memory Model — What Gets Persisted

MemoryModel

The unit of persistence is MemoryModel — a Pydantic model holding the conversation history of one session:

from amrita_core.types.memory import MemoryModel

memory = MemoryModel()  # empty history
memory.messages  # list[Message | ToolResult]

• messages — the conversation: Message entries (user / assistant) and ToolResult entries paired with their tool calls.
• Being a Pydantic model, it serializes with model_dump() and validates with model_validate() — exactly what a file/DB backend needs (see Data Backend).

The Lifecycle

1. Load — the workflow's LOAD_STATE node calls memory.load_memory(session_id) and stores the result in MemoryContext (chat._di_memory.memory).
2. Mutate — strategies append to SendMessageWrap; at the end (_post_runner) the assistant response is appended too, and the final list is written back into mem_ctx.memory.messages.
3. Commit — the COMMIT_MEMORY node calls memory.commit_memory(session_id, memory).

So the same session_id + backend combination determines what the next conversation loads — the framework only orchestrates the calls.

MemoryContext (DI)

The runtime memory lives in the MemoryContext DI slot:

chat._di_memory.memory  # MemoryModel | None — set after LOAD_STATE

Workflow nodes and strategies access it via type-matched injection (mem: MemoryContext).

Memory Summarization

LLMConfig.enable_memory_abstract + memory_abstract_threshold trigger summarization: when the prompt token count exceeds the threshold, older turns are replaced by a summary before the request is sent (see Tutorial 5 — Memory). The built-in step strategy additionally compresses history between Steps (see Step Loop).

StateContext (Legacy Accessor)

StateContext (session_id + memory + ability) still exists as a backward-compatible accessor: chat.state synthesizes one from the DI contexts, and LegacyBackend uses it as its in-process storage. New code should use the DI contexts (_di_memory, _di_ability, _di_session) directly.

Next

Data Management — back to the overview, or continue to Extensions & Integration.