#Recursive Memory Harness - RLM for Agentic Memory, give your Hermes agent superpowers.

1 messages · Page 1 of 1 (latest)

rotund topaz
#

An agentic harness that constrains models in three main ways:

Retrieval must follow a knowledge graph

Unresolved queries must recurse (Use recurision to create sub queires when intial results are not sufficient)

Each retrieval journey reshapes the graph (it learns from what is used and what isnt)

Smashes Mem0 on multi-hop retrieval with 0 infrastrature. Decentealsied and local for sovereignty

Now Hermes Agent Compatible: https://github.com/aayoawoyemi/Ori-Mnemos

been building an open source decentralized alternative to a lot of the memory systems that try to monetize your built memory. Something that is going to be exponentially more valuable. As agentic procedures continue to improve, we already have platforms where agents are able to trade knowledge between each other.

GitHub
GitHub - aayoawoyemi/Ori-Mnemos: Local-first persistent agentic mem...

Local-first persistent agentic memory powered by Recursive Memory Harness (RMH). Open source must win. - aayoawoyemi/Ori-Mnemos

#

link to RLM paper: https://arxiv.org/abs/2512.24601
Link to Paper Introducing RMH: https://orimnemos.com/rmh/

arXiv.org
Recursive Language Models

We study allowing large language models (LLMs) to process arbitrarily long prompts through the lens of inference-time scaling. We propose Recursive Language Models (RLMs), a general inference paradigm that treats long prompts as part of an external environment and allows the LLM to programmatically examine, decompose, and recursively call itself...

Ori Mnemos
Recursive Memory Harness — Ori Mnemos

Recursive Language Models changed how AI processes context. Recursive Memory Harness applies that same architecture to persistent memory — and retrieval quality compounds. Introducing RMH and its first implementation: Ori Mnemos.

hot sinew
#

I am very interested in this work @starro - I would love to discuss it with you.