Rome OS
Rome is an open-source, compounding agent operating system for recursive agents. It provides a guardrailed environment in which humans and agents collaborate, while tools, workflows, memory, apps and interfaces persist and improve across tasks.
At a Glance
- Individuals seeking a private personal AI agent
- Non-technical users who want a managed agent environment
- Developers and self-hosters
- Small teams and businesses automating recurring workflows
- +1 more
AI Tools by Rome OS
(1)Rome
Open Source Agent Environment
Discussions
No discussions yet
Be the first to start a discussion about Rome OS
Latest News
Rome Node v0.1.8 and Node Core v0.1.7 released with signer-JWT authentication to Rome Cloud.
Rome Node v0.1.7 released with a local append-only audit log on hosts.
Rome license update: the repository is MIT-licensed.
Initial public Rome repository commit recorded on GitHub.
Products & Services
Open-source agent OS for persistent agents, scheduled tasks, reusable workflows, custom apps and human-agent collaboration. It can be self-hosted with Docker or used through Rome Cloud.
Hosted private Rome environment provisioned and maintained for each user; offered in preview.
Installable, persistent applications combining purpose-built UI, agent reasoning, reusable workflows, typed actions, private data and schedules.
Catalog and distribution channel for Rome Apps, with app submission and review processes.
Market Position
Rome positions itself as a private, guardrailed agent OS that combines the convenience of a managed personal assistant with self-hosting, real installable apps, typed actions, persistent data and an app store. It explicitly compares with OpenClaw and Hermes Agent, and describes itself as an open-source alternative to Grok Bot and Meta's Muse; its differentiation is structural app capabilities and approval-oriented trust rather than only prompt/skill files.
Leadership
Founders
Yunfan Ye
Founder/building Rome OS; publicly describes himself as formerly of Google, Meta and Yutori AI. His LinkedIn search profile identifies Carnegie Mellon University graduate study in Information Networking.
Zihan (Zenus) Wang
AI researcher associated with Rome OS; public profiles describe work in agentic reinforcement learning, reasoning, foundation models and robotics, including affiliations with Northwestern University, NVIDIA and Meta.
Executive Team
Yunfan Ye
Founder / builder of Rome OS
Public profiles describe him as building Rome OS/Rome AI Lab and formerly working at Google, Meta and Yutori AI.
Zihan (Zenus) Wang
Co-founder / AI researcher
Public profiles describe research in agentic RL, reasoning, foundation models and robotics, with Northwestern University, NVIDIA and Meta affiliations.
Founding Story
Rome's founding thesis is that AI progress should scale not only models but also the environment around them: tools, workflows, memory and interfaces. The project aims to let agents construct and improve their own reusable capabilities under human guidance, so work produces both an immediate result and a durable capability for future work.
Business Model
Revenue Model
Open-source self-hosting is free; Rome monetizes managed private hosting through Rome Cloud subscriptions and offers custom Enterprise deployments.
Pricing Tiers
Run Rome on your own Mac; community Discord support.
Hosted by Rome for a single inbox and a few channels; 2GB memory and email support.
Hosted by Rome for heavier workloads and bigger memory; 4GB memory and priority email.
Hosted by Rome for heavy workloads and business use; 16GB memory and priority email.
Team-oriented deployment to the customer's server, with shared Slack support and SLA.
Target Markets
- Individuals seeking a private personal AI agent
- Non-technical users who want a managed agent environment
- Developers and self-hosters
- Small teams and businesses automating recurring workflows
- Organizations needing private, reviewable and guardrailed agent applications
- Email triage, organization and drafting replies
- Code review and resolving review comments
- Tracking prices, stocks and other topics with scheduled updates
- Customer interviews and research synthesis
- Building custom internal tools and workflows from an idea
- Long-running follow-through, reminders and recurring operational work