Working alongside Upstreet, we designed and developed a browser-based platform for creating, deploying, and managing AI agents without requiring developers to leave their workflow. Inspired by the simplicity of modern cloud platforms, the product provided a unified environment where teams could build agents, monitor deployments, configure integrations, and orchestrate complex agent ecosystems entirely from the browser.
Starting from a blank canvas, our Grayhatters helped shape the platform from initial product concept and UX direction through frontend and backend implementation, while also contributing to branding and launch materials. Users could create and configure both Mastra and ElizaOS agents directly within the platform, interact with them through integrated chat experiences, inspect logs and execution traces in real time, monitor container status, manage environment configuration, and deploy agents from reusable templates with a single click. The platform also supported collaborative multi-agent workflows, allowing specialized agents to work together on larger tasks and objectives.
A major focus of the platform was extensibility. Users could connect self-hosted MCP (Model Context Protocol) servers, enabling agents to securely access external tools, APIs, resources, and workflows through a standardized interface. The platform included a dedicated MCP registry, making it possible to connect hundreds of third-party services—including communication platforms, developer tools, and business systems—in just a few clicks. Combined with extensive observability, deployment controls, and agent orchestration capabilities, the result was a complete developer platform for building, deploying, and scaling production-ready AI agents.
