sno-ai/sno-station

Turn the agents you already use into a squad. Codex, Claude Code, Cursor, Hermes Agent and OpenClaw use one encrypted memory, pass work to each other, review each other's work, and dream every night to wake up a little better. Open source. Runs on your computer.

What it solves

Sno Station addresses the fragmentation and memory limitations of using multiple AI coding agents. It prevents work stoppages caused by rate limits, ensures consistent context across different tools, and automates the improvement of agent performance through a self-reflective "nightly loop."

How it works

Sno Station acts as a connective layer that links various AI coding agents (such as Claude Code, Cursor, and Codex) into a unified squad. It maintains a shared, encrypted memory on your local machine. When one agent hits a quota or rate limit, another can seamlessly pick up the task with full context. The system also features a "nightly loop" (REM) that reviews daily sessions, identifies recurring mistakes, and proposes skill updates for the agents, which you can then approve or reject.

Who it’s for

This project is for developers who use multiple AI coding agents simultaneously and want to synchronize their workflows, maintain persistent memory across different tools, and have their agents automatically learn from past mistakes.

Highlights

  • Unified Memory: Shares context and state across different agents like Cursor, Claude Code, and OpenClaw.
  • Automatic Handover: Allows agents to pass tasks to one another when one hits a rate limit.
  • Peer Review: Enables agents to review each other's work to catch errors that a single agent might miss.
  • Self-Improving Skills: Uses a nightly reflection process to analyze session logs and suggest improvements to agent behavior.
  • Local-First Privacy: Stores memory and skills locally on your machine with optional, user-controlled cloud synchronization.

Related

  • Project
    sno-ai/sno-station-skills

    Sno Station Skills is a collection of 31 self‑tested “skills” that you install into existing AI coding agents (Claude Code, Codex, etc.). The skills give agents team‑like habits – independent peer review, quota‑aware hand‑offs, health checks, messaging, and nightly retrospectives – so work can continue safely even when an agent runs out of quota or crashes. Installation is via a tiny `sno` command that auto‑updates and configures each skill for the agents you have.

  • Project
    codesstar/loci

    Loci is an open‑source, file‑based memory system that lets different AI agents read and write shared context stored locally as Markdown/JSON. It offers personal memory, task scheduling, project indexing, contact logs, a knowledge vault, and a local Node.js dashboard, all without requiring external databases.

  • Project
    hhyqhh/inno-agent

    Inno Agent is an open‑source personal learning AI built on the Pi coding‑agent SDK. It adds three memory layers (learner profile, wiki, session logs), a natural‑language scheduler, IM channel support (Feishu/WeChat), and an in‑browser practice lab. Available as an Electron desktop app, a React web UI, or a terminal CLI, it works with any OpenAI‑ or Anthropic‑compatible model and can be extended via a skill hub. The project is MIT‑licensed and includes Docker, sandboxing, and a session‑replay showcase.

  • Project
    TencentCloud/TencentDB-Agent-Memory

    TencentDB Agent Memory is an open‑source memory hub for LLM agents. It captures conversations, extracts reusable skills, builds a Wiki from docs, and indexes code into a CodeGraph. A central UI (Memory Hub) lets teams manage assets, set fine‑grained ACLs, and bind specific memory pieces to individual agents. Agents talk to a uniform proxy, so any framework (Claude Code, DeepSeek Harness, OpenClaw, etc.) can use the same memory without code changes. Benchmarks show a 59 % boost in persona retention. The repo provides one‑click Docker deployment, migration tools, and a roadmap for zero‑config cold starts and more integrations.

  • Project
    solo-agent/solo

    An open-source, local-first workspace that allows humans and multiple AI coding agents to collaborate via channels, threaded task boards, and shared memory.