VarOps
VarOps helps established companies become AI-native by building Resident AI: organizational intelligence that operates inside the company's existing environment, understands how the business actually runs, works under company authority, and remains owned by the company. Its stated principle is that a company should own its knowledge and rent the thinking.
At a Glance
- Mid-market companies
- Established companies and enterprise leadership teams
- Organizations requiring data, model and infrastructure sovereignty
- Builders, operators, founders and CEOs responsible for practical AI adoption
- +1 more
AI Tools by VarOps
(1)actp
Remote Coding Agent Transport Bridge
Discussions
No discussions yet
Be the first to start a discussion about VarOps
Latest News
VarOps published its formal Resident AI definition and five-condition test.
Old School / New Tech published Your agent isn’t stupid. It’s blind., describing lessons from a year of building organizational memory at VarOps.
VarOps/actp released v0.1.0, the first ACTP release, with CLI, desktop builds and remote ACP-agent functionality.
Ran Aroussi announced the public launch of VarOps, its Resident AI thesis, Acropolis research, transformation work and magazine.
Products & Services
VarOps's operating model and category for AI that lives inside an organization's environment, understands institutional context, works under organizational authority and governance, and keeps the company's knowledge separate from any model vendor.
A company-knowledge and organizational-memory system that observes approved sources, maintains a client-owned, versioned and exportable operational model, governs identity and permissions, and supplies bounded evidence to reasoning and execution components. It is designed to run inside infrastructure controlled by the client.
Founder-led, paid engagements for mid-market companies following Observe, Understand, Build, Run, and Compound: diagnose the operation, model how work really happens, build fixed automations or bounded agents, deploy them in production, and improve them over time.
Open-source ACP-over-HTTP software from the VarOps GitHub organization. It lets an ACP client, app, bot, or editor use coding agents such as Claude Code, Codex, Droid, or Pi on another machine through a bridge and server; it includes CLI, desktop/tray app, dashboard, tokens, Docker, and Cloudflare tunnel support.
Market Position
VarOps positions Resident AI against bolted-on chatbots, copilots, generic agent platforms, vendor-hosted RAG and conventional AI consulting. Its differentiation is a client-owned operational model running inside the company's environment, with explicit boundaries between observation, memory, authority and execution. For ACTP, the relevant alternatives are local-only ACP clients and agent-remote-access/orchestration tools; ACTP emphasizes protocol compatibility, no client or agent changes, resilient reconnects, scoped per-client tokens, and no required central server.
Leadership
Founders
Ran Aroussi
Founder of VarOps and CEO of Automaze. He describes 35+ years spanning software engineering, entrepreneurship, performance marketing, ad-tech infrastructure, and operating businesses; he created yfinance and other widely used open-source software, wrote books on production AI, and hosts the Old School / New Tech podcast.
Executive Team
Ran Aroussi
Founder
Founder of VarOps and CEO of Automaze; software engineer and entrepreneur with 35+ years of experience, creator of yfinance and other open-source tools, author of production-AI books, and host of Old School / New Tech.
Founding Story
VarOps was developed over roughly six months after Aroussi observed that AI systems can reason well but generally cannot see the real operating model of a company, which is distributed across email, meetings, chats, systems, and people's tacit knowledge. The initial vision was to build a company-owned layer of living organizational understanding between a business and changing AI models, then use it to govern and execute work without forcing employees into new workflows.
Business Model
Revenue Model
VarOps says it does not sell software as a product line. It earns revenue from paid, time-boxed diagnosis and transformation engagements for mid-market companies, in which it observes the operation, builds and deploys resident-AI infrastructure, and supports production operation. The company also says workshops are planned as paid deeper work; its magazine and newsletters are free.
Pricing Tiers
A fixed 30-day paid engagement intended to determine what level of AI a company actually needs.
A transformation phase that observes approved systems and maintains a continuously updated model of how the organization operates.
A production phase in which VarOps builds, governs and runs agents inside the tools the company already uses.
Target Markets
- Mid-market companies
- Established companies and enterprise leadership teams
- Organizations requiring data, model and infrastructure sovereignty
- Builders, operators, founders and CEOs responsible for practical AI adoption
- Teams running coding agents on separate workstations, servers or VPSs
- Building an operational model of an established company's real workflows, decisions, ownership and dependencies
- Enterprise knowledge management and organizational memory
- Running AI on company-controlled infrastructure where data sovereignty matters
- Automating stable workflows and deploying bounded agents where judgment is required
- Governed execution in Slack, Microsoft Teams and existing operational tools
- Diagnosing whether a business needs AI before committing to a rollout