# pxmo / Proply

> Helped evolve a collaborative proposal platform from core editor and rendering systems through secure acceptance, product AI and agent-driven document operations.

- **Role:** Full-Stack Product & Agent Platform
- **Category:** Professional work
- **Human-readable case study:** https://lutherminshull.com/projects/pxmo-proply

## Context

pxmo / Proply is a digital proposal platform with collaborative documents, media-rich authoring, rendered customer outputs, secure acceptance and tenant-aware workflows.

## Luther's role

- Contributed to the product from its formative dashboard, template, inventory and team workflows.
- Worked inside the proposal editor, media systems and customer-facing renderer rather than only around the application shell.
- Delivered secure acceptance, output reliability and later product-AI features.
- Built and verified an MCP platform for structured, authorized agent operations over proposal content.

## What Luther built

- Proposal and template workflows, team and profile management, inventory lists and collaborative item editing.
- Editor modules and media discovery across Unsplash, Pexels and GIPHY, including browser and rendered output behavior.
- Configurable signature-verification levels, custom verification codes and secure proposal acceptance.
- PDF pagination, responsive content and media-rendering reliability improvements.
- AI proposal summaries and strategy analysis delivered through persisted, streamed product workflows.
- Delegated OAuth authorization, tenant and environment isolation, semantic proposal operations and audit and idempotency controls for MCP.
- Snapshot-bound proposal previews and Docker-based rendering infrastructure.

## How it works

- React and Laravel product surfaces connected to shared editor state, WebSocket services and Yjs documents.
- Proposal behavior kept consistent across editable, read-only, mobile and PDF output surfaces.
- MCP tools delegate through the application authorization layer rather than accessing raw tenant data directly.
- Semantic document operations expose structured proposal views while raw collaborative document state remains internal.

## Outcomes

- Supported the platform's evolution across core product, authoring, acceptance, rendering and AI capabilities.
- Established a verified path for agents to interact with proposal content through explicit authorization and environment boundaries.
- Improved document reliability across interactive and generated customer-facing outputs.

## Technologies

React, Next.js, Laravel, PHP, Yjs, MCP, OAuth 2.0, Docker

## External link

- [Visit pxmo](https://pxmo.com)

## Contact

- [Discuss an opportunity](https://cal.com/lutherminshull)
- [Email Luther](mailto:luther@lutherminshull.com)
- [Return to the portfolio overview](https://lutherminshull.com/index.md)

## Scope note

This case study is limited to public and non-sensitive implementation details. Client-confidential information and internal data are not included.