ant_ai.memory.protocol
Memory
pydantic-model
Bases: Tool
Base class for an agent's long-term memory, kept across conversations.
A memory is a tool namespace: its public methods are the tools the model
calls to recall and save what it knows, registered as <ClassName>_<method>
when the memory is set as Agent(memory=...). Each backend picks the tools
that fit its storage: Mem0Memory searches and adds facts, FileMemory reads
and writes one Markdown file.
Example
from ant_ai.core.types import InvocationContext
from ant_ai.memory import Memory
class NotesMemory(Memory):
notes: list[str] = []
def recall(self, ctx: InvocationContext | None = None) -> list[str]:
# The docstring of a method is the tool description the model reads.
return self.notes
def save(self, note: str, ctx: InvocationContext | None = None) -> str:
self.notes.append(note)
return "Saved."
Notes
A method that declares ctx receives the current InvocationContext
(e.g. ctx.user_id, to keep each user's memories apart); it is never
shown to the model.
Show JSON schema:
{
"description": "Base class for an agent's long-term memory, kept across conversations.\n\nA memory is a tool namespace: its public methods are the tools the model\ncalls to recall and save what it knows, registered as `<ClassName>_<method>`\nwhen the memory is set as `Agent(memory=...)`. Each backend picks the tools\nthat fit its storage: `Mem0Memory` searches and adds facts, `FileMemory` reads\nand writes one Markdown file.\n\nExample:\n ```python\n from ant_ai.core.types import InvocationContext\n from ant_ai.memory import Memory\n\n\n class NotesMemory(Memory):\n notes: list[str] = []\n\n def recall(self, ctx: InvocationContext | None = None) -> list[str]:\n # The docstring of a method is the tool description the model reads.\n return self.notes\n\n def save(self, note: str, ctx: InvocationContext | None = None) -> str:\n self.notes.append(note)\n return \"Saved.\"\n ```\n\nNotes:\n A method that declares `ctx` receives the current `InvocationContext`\n (e.g. `ctx.user_id`, to keep each user's memories apart); it is never\n shown to the model.",
"properties": {
"name": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Tool name.",
"title": "Name"
},
"description": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Tool description. Is used by the LLM to decide whether to call or not the specific tool.",
"title": "Description"
},
"parameters": {
"anyOf": [
{
"additionalProperties": true,
"type": "object"
},
{
"type": "null"
}
],
"default": null,
"description": "The parameters needed by the tool. This is a self-constructed field.",
"title": "Parameters"
}
},
"title": "Memory",
"type": "object"
}
Fields:
-
name(str | None) -
description(str | None) -
parameters(dict[str, Any] | None) -
__namespace_methods__(list[str])
Validators:
-
_set_defaults
Source code in src/ant_ai/memory/protocol.py
6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 | |