Skip to content

Project Roadmap

Jump to bottom
Marcus Sanatan edited this page Oct 17, 2025 · 6 revisions

MCP for Unity - Project Roadmap

Welcome to the MCP for Unity roadmap! This document outlines our high-level goals, priorities, and planned features for the project. It's intended to give users and contributors insight into where the project is heading, enabling seamless communication between the Unity Editor and LLM tools like Claude Desktop and Cursor via the Model Context Protocol (MCP).

Please Note: This roadmap is a living document and reflects our current plans and priorities. These are subject to change based on community feedback, unforeseen challenges, or shifts in project direction. Timelines are estimates and not guarantees. The dates on linked issues/PRs may reflect when they were opened, not necessarily the target completion date.

Want to contribute or discuss the roadmap? We'd love your input! Please see the How to Contribute or Provide Feedback section below.

Table of Contents

Overall Vision & Goals

Our primary goal for MCP for Unity is to provide a robust, flexible, and easy-to-use bridge and MCP server enabling powerful interactions between the Unity Editor and other MCP clients.

Key objectives driving our development include:

  • Goal 1: Improve Distribution and Setup: We're available on GitHub and OpenUPM, we're working towards being available on the Unity Asset Store.
  • Goal 2: Enhance Stability & Performance: Ensure reliable communication and efficient handling of commands, especially in large projects.
  • Goal 3: Expand Features & Integrations: Add more Tools, leverage Resources and Prompts, support more Unity features (like HDRP), and integrate with more AI services/tools.
  • Goal 4: Improve User & Developer Experience: Enhance documentation, provide better feedback/error handling, and support community contributions like localization.
  • Goal 5: Increase Maintainability & Testability: Refine the architecture and add tests to ensure long-term health.

Current Focus / Next Release (v7)

Legend: ✨ Feature, 🐛 Bug Fix, 📈 Improvement, 📄 Docs, 🧪 Tests, 🧹 Tech Debt/Refactoring, 🏗️ Architecture

We are currently focused on becoming available on the Unity Asset Store. On the Asset Store, we cannot include executables or redirect users to get them. As such, we need to make the plugin as self-contained as possible. That means, the user cannot received errors for not having dependencies like Git, Python and uv, and they needs to be some measure of working functionality without them.

We've submitted a few times but Unity has rejected us. We'll do our best to see it through! This is currently being tracked in #285

Target: Aiming for completion in early November 2025. Of course, this depends on Unity as much as it depends on us.

Key items being actively worked on or planned for this cycle:

  • 🧹 Return typed classes instead of a Python dictionaries in the MCP server
  • ✨ Prepare and upload to Unity Asset Store #6
  • ✨ [COMMUNITY] Support GitHub Copilot in Jetbrains #312
  • ✨ [COMMUNITY] Custom tools #309
  • 🐛 Fix connections to Codex on Windows #315, #319, related #311
  • 📈 Paginate and optimize scene hierarchy (and make it a resource) #317
  • 📈 Paginate and optimize read console (and make it a resource) #302
  • 🐛 Fix testing suite #300
  • 📈 Address shader warning #290

Short-Term Plans

  • 🧹 Use less static classes and create more services
  • 🧹 Use typed parameters/classes instead of a JSON object in the tool parameters
  • 📈 Utilize resources and prompts so MCP clients can perform better #287

Mid-Term Plans

These are items we aim to tackle further out. Details and priorities are less defined.

  • ✨ Explore Runtime MCP Operation (#58).
  • 🏗️ Dependency Injection to improve testability
  • 📈 Incorporate screenshots into LLM decision making
  • ✨ Explore adding GenAI plugins for 2D and 3D assets
  • ✨ Fully support TCP and HTTP connections to the MCP Server

Long-Term Ideas & Future Directions

This section captures bigger ideas or major features that are further down the line (>6-9 months) or require significant research/design.

  • 📈 Create more auto-configurations to support for additional MCP clients.
  • ✨ Add more play mode functionality - support MCP during runtime with custom tools would allow for LLMs to interact with user created games/experiences
  • ✨ Add HDRP support (#27).

Maybe / Icebox / Backlog

These are ideas that have been suggested or considered but are not currently planned for active development. They might be revisited later.

  • Visual scripting integration (e.g., Bolt/PlayMaker).
  • 🧪 Test coverage for Tools, networking and ideally end-to-end

Recently Completed

For details on past releases and completed work, please see:

How to Contribute or Provide Feedback

Your feedback and contributions are crucial! Here's how you can get involved:

  1. Discuss Ideas: Use GitHub Discussions or open an issue with the enhancement or discussion label to discuss roadmap items or propose new ideas.
  2. Request Features: For specific, well-defined feature requests, please open a new issue using the "Feature Request" template. Check existing issues first!
  3. Report Bugs: If you find a bug, please open a bug report. Provide clear steps to reproduce.
  4. Contribute Code/Docs: Check our README.md guide for details. Look for issues tagged help wanted or good first issue. Review open Pull Requests.
  5. Comment on Issues/PRs: Provide feedback directly on the issues and pull requests linked in the roadmap sections above.

Disclaimer

This roadmap provides a high-level overview of potential future direction for MCP for Unity. It is not a commitment or guarantee. Priorities and timelines may change significantly based on various factors including (but not limited to) community feedback, resource availability, technical challenges, and strategic shifts. We will strive to keep this document updated but please refer to specific GitHub Issues and Milestones for the most granular, up-to-date status.

Clone this wiki locally