Smart Node opens its AI front office platform to small businesses for free

Smart Node, a self-hosted AI front office system for small businesses, is now open source under AGPL-3.0 and available at no license cost. The software runs on the owner’s own machine, keeps customer data local, and requires approval before any agent can make a commitment to a customer.

Why it matters: - Small businesses can automate front-office work without paying a licence fee or handing data to a third-party cloud. - Smart Node keeps customer records and conversations on the owner’s machine, which can reduce privacy and control concerns. - The approval step for quotes and bookings gives owners direct oversight before any customer commitment goes out.

What happened: - Smart Node, built by Chicago software engineer Dmytro Klymentiev, is now open source under the AGPL-3.0 licence. - The current release is version 1.14.1 and is available on GitHub. - The platform is designed to run a crew of AI agents for the front office on a business owner’s own machine.

The details: - The agents handle routine front-office work including answering messages, following up on leads, booking appointments and keeping paperwork organized. - Any action that commits a business to a customer, including a quote or a booking, waits for the owner’s approval before being sent. - Smart Node uses separate agent teams for each business instead of a single general chatbot. - The system includes specialists for the front desk, sales and back office. - Plain-language rules limit what the agents can do. - Workflows divide tasks so one agent completes a step and passes the work to the next. - Smart Node ships 50 ready-made crews for trades and service businesses, including dental clinics, restaurants, plumbers and accounting firms. - The platform also includes 95 roles across 17 departments. - A setup assistant can adapt a crew to other types of business. - The system can run on an office computer, a home server or a small VPS. - A single panel shows every action the agents take in real time. - Model calls route through a gateway to Claude, GPT or Gemini. - A local model is available on request. - The agents run on Flint, Smart Node’s open-source agent. - Claude Code can connect through an adapter. - Smart Node runs on Debian or Ubuntu and needs at least 2 GB of memory and 10 GB of disk. - Businesses can also ask the Smart Node team to handle installation, adapt a crew to their trade and connect existing tools. - The website is Smart Node and the source code is available on GitHub.

Between the lines: - The product is aimed at small businesses that want AI automation but still need human control over customer-facing decisions. - Self-hosting positions Smart Node as an alternative to cloud-based AI front-office tools. - The combination of ready-made crews and customization suggests the product is trying to balance speed of setup with industry-specific workflows.

What's next: - Smart Node will likely focus on adoption by small businesses that want to self-host rather than subscribe to a managed AI service. - Businesses that do not want to install the system themselves can use the team’s setup and integration help. - The open-source release may invite outside contributions and broader deployment beyond the initial trade templates.

The bottom line: - Smart Node packages agentic AI front-office automation into a free, self-hosted system that keeps business owners in the approval loop.