Most software today is designed for humans. A person opens an app, navigates menus, fills out forms, clicks buttons. The entire interface assumes a human is on the other end.
We believe that assumption is about to change—fundamentally and permanently. Over the next several years, the majority of interactions with software like TIDY will not come from humans clicking through a UI. They will come from AI agents acting on behalf of humans—personal assistants, property management agents, home automation systems, and general-purpose agents that handle the logistics of daily life.
We are building accordingly. Today we are launching TIDY's new MCP server and general-purpose API, designed from the ground up for the agentic era.
What We Are Launching
Two things, both aimed at making TIDY the easiest way for any AI agent to handle cleaning and maintenance.
A new MCP server. Model Context Protocol is the open standard for connecting AI agents to external tools and services. Any MCP-compatible agent—Claude, and a growing number of others—can now connect to TIDY's MCP server and make cleaning and maintenance requests directly. The agent does not need to know our API schema in advance. It discovers TIDY's capabilities through MCP's tool discovery protocol, understands what actions are available, and calls them as needed.
A new general-purpose API. We have updated and expanded our REST API with endpoints designed specifically for agent consumption. These endpoints accept natural language alongside structured data, so an agent can describe a problem in plain English—"the kitchen faucet is dripping and there's water on the floor"—and TIDY's AI interprets the request, routes it to the right service category, and returns a structured response the agent can act on.
The core idea is simple: an agent can call into TIDY with almost any cleaning or maintenance request, and we can respond back—with actions, status updates, options, or clarifying questions. TIDY becomes a service that agents can use the same way a human would use a property manager: describe what you need, and we handle it.
Why Agents Change Everything
Consider how property maintenance works today. A homeowner notices a problem. They open an app or make a phone call. They describe the issue, wait for a response, compare options, confirm a time, and follow up to make sure the work was done. Each step requires the human's attention.
Now consider what happens when an AI agent handles this. The homeowner tells their assistant—in conversation, by text, or through any interface—"the kitchen faucet is dripping." The agent takes it from there. It files the request with TIDY. TIDY's system classifies the issue, checks the property's digital twin for relevant context (what kind of faucet, when it was last serviced, which plumber has the best track record for this type of repair), proposes options, and sends them back to the agent. The agent confirms on behalf of the owner, or asks a clarifying question if needed. The repair is scheduled, the pro is dispatched, and the agent follows up when the job is complete.
The owner's total involvement: one sentence.
Or take a more proactive scenario. A general-purpose home management agent notices it's been three months since the last deep clean. Based on the owner's preferences, it initiates a request through TIDY—no prompt from the owner at all. TIDY schedules the cleaning, the pro confirms, the job gets done. The owner gets a notification that their home was cleaned. That is it.
The interface shifts from UI to conversation. From human-driven to agent-driven. TIDY needs to be on the other end of that conversation, and with our new MCP server and API, we are.
How It Works
MCP Server
MCP is an open protocol that lets AI agents discover and use external tools. When an agent connects to TIDY's MCP server, it receives a list of available tools—things like requesting a cleaning, scheduling maintenance, checking job status, or getting a quote. The agent can then call these tools as part of a conversation with the user.
From the agent's perspective, TIDY is just another tool in its toolkit. The agent does not need custom integration code or knowledge of our internal data model. It discovers what TIDY can do, and uses those capabilities as needed.
General-Purpose API
Our updated REST API is designed for flexibility. Traditional APIs require callers to know the exact endpoint, parameters, and data format. That works for human developers building integrations, but agents operate differently. They reason about problems in natural language and need interfaces that can meet them where they are.
TIDY's new API endpoints accept both structured requests and natural language descriptions. An agent can send a fully structured request if it has the data, or a plain English description if it does not. TIDY's AI layer handles the interpretation either way—classifying the request, extracting relevant details, and routing it to the right workflow.
Responses are structured and machine-readable, designed for agents to parse and act on without ambiguity. Status codes, action options, estimated timelines, and next steps are all returned in a format that agents can reason about programmatically.
Webhooks for Agent Awareness
Agents need to stay informed. Our webhook system notifies agents when events happen—a job is completed, an issue is detected, a pro confirms or reschedules. This allows agents to proactively update their users or take follow-up actions without polling.
Full documentation is available in our developer docs.
Why We Are Building for Agents First
This is not a side project or a nice-to-have integration. We believe agents will be the primary way people interact with services like TIDY within the next few years.
Think about the trajectory. Today, most people manage their homes reactively. Something breaks, they deal with it. Cleaning gets scheduled when someone remembers to schedule it. Maintenance gets done when it becomes urgent. The friction of coordinating these things means they are often deferred, forgotten, or handled poorly.
Agents eliminate that friction entirely. When your AI assistant can handle a maintenance request in the time it takes you to mention it, the entire model shifts from reactive to proactive. Homes get maintained better because the cost of coordinating maintenance drops to near zero.
We have spent 12 years building the infrastructure to coordinate cleaning and maintenance at scale—digital twins of properties, AI models trained on data from over 100,000 homes, a vetted network of service providers, predictive scheduling, quality verification. All of that infrastructure now sits behind an interface that any agent can use.
The hard part was never the interface. The hard part was building the operational backbone that makes it possible to say "fix my faucet" and have it actually happen, reliably, at scale. That backbone exists. Now we are opening it up to every agent that wants to use it.
What This Means
For property owners and homeowners: Your AI assistant—whatever form it takes—can now manage your cleaning and maintenance through TIDY. Describe a problem in conversation and it gets handled. Set preferences once and let your agent manage the rest proactively.
For developers building agents: If your agent needs to handle anything related to cleaning or maintenance, TIDY is the backend. Connect via MCP for the simplest integration, or use our REST API for more control. Either way, you get access to TIDY's full operational infrastructure—scheduling, pro dispatch, quality verification, digital twins—without building any of it yourself.
For the industry: The coordination layer is becoming invisible. The future is not humans managing software that manages properties. It is agents managing properties through services like TIDY, with humans involved only when they want to be. We are making sure TIDY is ready for that future.
Start Building with TIDY
Connect your agent to TIDY's MCP server or explore our API to add cleaning and maintenance capabilities to any application.