calpaterson/memoryfield-spec

A model-independent format for storing and sharing memories for use by AI agents and humans

stars

commits

Sep 5, 2026

updated

README

memoryfield spec

This repo contains the RFC-style specification for the "memoryfield" AI agent memory format.

memoryfields are a vastly simpler way of giving AI agents memories. The memoryfield format is based on standard file formats with wide support that both LLM agents and humans can easily understand.

For an introduction to the format, read Agent memory as a file format.

For the RFC-style specification, see SPEC.md.

For the Agent Skill implementation, see the repo calpaterson/memoryfield-skill.

Contributors

calpaterson

7 commits

See what people are saying

Hacker News (1)

Agent Memory as a File Format

0 points · Aug 31, 2026

Similar repos

moorcheh-ai/memanto

Memory that AI Agents Love!

1,861

huggingface/funes-memory

Dataset

agentscope-ai/ReMe

ReMe: Memory Management Kit for Agents - Remember Me, Refine Me.

3,445

LightningCreeper/MIA

Model

BAI-LAB/MemoryOS

[EMNLP 2025 Oral] MemoryOS is designed to provide a memory operating system for personalized AI…

1,573

ssanvi-builds/seahorse

Open standard for persistent, self-evolving LLM agent memory (open-core, Apache-2.0).

zilliztech/memsearch

A persistent, unified memory layer for all your AI agents (e.g. Claude Code, Codex, DSH), backed by…

2,588

ViktorAxelsen/MemSkill

MemSkill: Learning and Evolving Memory Skills for Self-Evolving Agents

572