Our work
Help an idea become
something people can use.
We are connecting AI creation with open publishing, free public infrastructure, and practical guidance for the person behind the idea.
Open standards
The Deploy Context Protocol (DCP) is intended to give AI agents and other tools a common way to publish and manage applications. Instead of asking a person to work through every hosting step, an agent could submit the project and receive its public address and ownership information.
The standard is being designed around publishing a project, checking its state, reviewing problems, updating it, and giving its human creator a clear path to control. Export and portability are also design goals.
Why an open standard?
People use different agents and development tools. A shared interface can make publishing available across those tools and allow other providers to implement the same approach.
Public infrastructure
We are preparing dcp.run as a public implementation of the Deploy Context Protocol (DCP). The intended experience begins with something a person and their agent have created.
A person brings an idea
They work with their agent to build a useful first version.
The agent takes it online
The service returns a public web address and a route to ownership.
The person stays in control
They claim the project, share it, and keep improving it.
Free publishing is the goal
We plan to offer the public service at no cost to creators at launch, with no paid plans and fair use limits that make shared resources sustainable. This describes DCP hosting; any charges for the AI tool a person chooses are separate.
Supported project types, resource limits, retention, and ownership rules need to be published before launch. We are preparing isolation, backups, network controls, and responses to abuse as part of that work.
Education
Our education program invites people to explore what they could create with AI. Start with an interest, an inspiration, or something you would love to share, then discover how an agent can help bring it to life.
The guides cover shaping a first project, giving an agent a clear brief, improving the result through feedback, checking it with real examples, sharing it with a community, and retaining ownership.
The guides are open to people and agents, with no DCP account required. They offer a shared starting point for exploring ideas and learning through creation.
Research
We intend to study the barriers between a person’s idea and a useful public project. Our first questions are practical:
- How can someone with no publishing experience describe what they want and assess what an agent creates?
- Where does an agent need to explain a decision or return control to the human?
- How can a public service make deployment simple while keeping ownership understandable?
- Which limits and safety measures preserve free access without making small projects difficult to share?
- What helps a creator keep improving a tool after its first users try it?
No research papers or study results have been published by the project yet. Future reports should explain their methods, limitations, and findings.