The Story Behind UseContext

Building the Context Cloud for AI Agents
By Nicholas Underwood
This was half-written by Claude Code and needs to be edited for accuracy and authenticity. It's on the TODO list.

The Vision

A crowd-curated context cloud — the central repository for collections of contexts on a growing variety of topics — fully accessible via MCP, API and the web. UseContext is the one-stop shop for all your (and more importantly your agent's) context needs.

The Genesis: From Context Garden to UseContext

The idea first sparked in early 2025 under the name "Context Garden". The vision was simple yet potentially powerful: create a crowd-curated context cloud that would serve as the central repository for collections of contexts on virtually any topic.

In our rapidly evolving AI landscape, both human developers and AI agents are constantly seeking relevant, high-quality context to inform their work. UseContext was designed to bridge this gap, making contextual intelligence fully accessible via MCP (Model Context Protocol), API, and web interfaces.

The Hackathon Sprint

The project transformed from concept to reality during a large hackathon hosted by bolt.new in June 2025. What followed was an intense "mad dash" of development on nights and weekends to develop a mostly production ready system. Little did I know that judging did not seem to take into account the working state of the application.

During the competition, I discovered a subcategory prize sponsored by a domain services company. To qualify, participants needed to use a .online domain. This requirement led to the acquisition of usecontext.online and, ultimately, the evolution from "Context Garden" to "UseContext."

The name "UseContext" has grown on me considerably. It's direct, functional, and immediately communicates the platform's purpose.

Technical Vision

UseContext was architected as a multi-access platform, recognizing that different users and systems would need different ways to interact with contextual data. The three primary interfaces were:

MCP Integration: Native support for the Model Context Protocol, enabling AI agents to seamlessly discover and consume relevant contexts during their operations.

RESTful API: A robust API layer for programmatic access, allowing developers to integrate contextual intelligence into their applications and workflows.

Web Interface: A human-friendly web platform for browsing, contributing, and managing context collections.

The Judging Period Pause

Following the hackathon submission, there was an embargo period during July 2025 while the judging process took place. This enforced pause provided valuable time for reflection on the project's potential and the broader implications of contextual intelligence platforms.

I have become increasingly convinced that UseContext might have "some legs and a path to real usage." I'm also flirting with the idea of positioning it as a "context marketplace" as a monetization strategy.

Current Status and Future

As of August 2025, UseContext is fully functional, but needs polish and improvements before marketing.