rho  by mikeyobrien

Persistent AI agent with local memory and self-awareness

Created 3 weeks ago

New!

279 stars

Top 93.3% on SourcePulse

GitHubView on GitHub
Project Summary

Summary Rho is an AI agent designed for continuous operation, session-to-session memory recall, and autonomous check-ins, prioritizing user data privacy by keeping all memories and data on-device. It targets users on macOS, Linux, and Android requiring a persistent, self-managing AI assistant without cloud reliance for core functionality. Key benefits include local data ownership, control, and flexibility to integrate with user-chosen LLM providers.

How It Works The agent uses a periodic "heartbeat" mechanism for configurable check-ins, processing tasks and reminders stored in its "brain." This brain, persisted as brain.jsonl, centralizes identity, learnings, and context for session continuity. Rho employs a modular architecture with "skills" (runbooks) and "extensions" (TypeScript code) to extend capabilities, enabling local data processing and user-defined LLM integration.

Quick Start & Requirements Installation is primarily via npm install -g @rhobot-dev/rho or cloning the repo and running ./install.sh. Prerequisites include Node.js (18+), tmux, and git. Rho supports macOS, Linux, Android (via Termux), and iPhone/iPad through SSH. The installer checks for missing dependencies.

Highlighted Details

  • Local-First Memory: All agent data stored locally in ~/.rho/brain/brain.jsonl ensures privacy and ownership.
  • Autonomous Operation: Features a periodic "heartbeat" for self-initiated check-ins and task execution.
  • Cross-Platform Skills: Offers system skills (notifications, clipboard, TTS) and platform-specific capabilities, notably extensive Android features (SMS, STT, media, Tasker).
  • Agent Email: Provides a name@rhobot.dev email address for direct communication.
  • Web UI: Browser-based interface for chat, memory, tasks, and configuration.
  • Extensibility: Differentiates markdown "skills" from TypeScript "extensions."

Maintenance & Community Specific details on maintainers, sponsorships, or community channels are not explicitly provided in the README.

Licensing & Compatibility The README does not explicitly state the software license, a critical factor for adoption. Compatibility is confirmed for macOS, Linux, and Android, with SSH setup for iOS/iPadOS.

Limitations & Caveats The absence of a declared license is a critical adoption blocker. iOS/iPadOS setup requires SSH. Some features, like the Telegram adapter, are noted as MVP and polling-first. Users must provide their own LLM provider.

Health Check
Last Commit

21 hours ago

Responsiveness

Inactive

Pull Requests (30d)
17
Issues (30d)
11
Star History
286 stars in the last 22 days

Explore Similar Projects

Feedback? Help us improve.