ToolExperimentalA2AMCPMODEL-CONTEXT-PROTOCOL

emem

Name: emem
Rating: 3.0 (56 reviews)
Author: Vortx-AI

by Vortx-AI

Rust long-term memory store for agents with A2A/MCP integration

Rust

Updated Sep 4, 2026

Stars

Forks

Commits/Week

514

Commits/Month

View on GitHub

Overview

Provides a Rust-based external memory store agents can carry outside their process to persist beliefs, observations, and world models. Implements long-term memory primitives that integrate with A2A/MCP-style workflows so multiple agents or services can read/write and evolve the same memory over time. Designed for long-running agents and earth-observation workflows, it focuses on structured memory objects, time/versioning, and retrieval for subsequent reasoning. Tree of Thoughts Pattern MCP-style workflows

See related protocols

Standards this tool supports

Key Benefits

As agents operate across sessions and collaborate, a shared, durable memory becomes essential for tracking behavior, failures, and reputation over time. Externalizing memory makes agent track records auditable and enables continuous evaluation of agent reliability across interactions. This repo matters because persistent memory is a foundational signal for multi-agent trust, informing evaluation, governance, and reputation systems. Evaluation-Driven Development (EDDOps)

Ideal For

Developers building long-running or distributed agents who need durable, shareable memory for provenance, world models, and agent track records. Agent-to-Agent Protocol (A2A)

Applications

Persisting agent observations and beliefs across sessions for continuous evaluation
Sharing world models between specialist agents in multi-agent workflows
Recording agent decisions and provenance to feed reputation or RepKit-style scoring
Supporting long-running earth-observation agents that need versioned historical context