Model or dataset
prajwalshettydev/UnrealGenAISupport avatar
prajwalshettydev/UnrealGenAISupport

The free plugin's readme says its own MCP support is not being developed

Unreal Engine plugin for LLM/GenAI models & MCP UE5 server. OpenAI GPT-5, Deepseek R1, Claude Opus/Sonnet, Gemini 3, Grok 4, Alibaba Qwen, Kimi, ElevenLabs TTS, Inworld, OpenRouter, Groq, GLM, Ollama, Local, Meshy, Tripo, Hunyuan3D, Rodin, fal, Dashscope, Seedream. NPC AI, agentic, chat, 3D gen, TTS, multimodal, image gen. UnrealMCP/UnrealClaude

686 stars111 forksC++MIT

At a glance

What is it?
UnrealGenAISupport is a free Unreal Engine plugin that wraps hosted model APIs so a project does not build its own integration layer. Its model list trails the paid plugins by a generation or two, its local-model support lives in another repository, and the cross-platform testing it lacks is sold separately.
Who is it for?
UnrealGenAISupport suits a team that wants hosted chat and reasoning models wired into an Unreal project this week without writing the integration layer, and that is content to keep the plugin's model identifiers updated by hand as vendors rename them. It does not suit a team that needs text-to-3D, speech, or an MCP server that will still be maintained, since none of those are in the free plugin and the MCP code is documented as unmaintained.
Can I use it commercially?
Yes. MIT is a permissive licence: you can use, modify and sell software built on it, as long as you keep its copyright and licence notices.
Is it still maintained?
Activity is slowing. The repository last received commits 6 months ago.
What is it written in?
Mainly C++, according to GitHub's language statistics.

Answers come from the project's GitHub data, last synced on November 15, 2026, and from our analysis. They are not legal advice.

From daily snapshots

Star history

  • prajwalshettydev/UnrealGenAISupport686 ★ · +42
Star history
Repository2026-09-112026-11-15
prajwalshettydev/UnrealGenAISupport644686

42 Stars in the last 65 days, measured from the daily GitHub snapshot. Compare with other projects

Editorial analysis

The readme says its own MCP support is not being developed

The most important sentence in the document sits in a blockquote at the end of the MCP section, and it says the MCP support in this free plugin is not being actively developed, with the listed features reflecting the current state. That sits awkwardly beside two other facts in the same repository. The project description advertises an MCP server for Unreal Engine 5, and the first of the two usage examples is an MCP one, where Claude spawns scene objects, changes their transforms and materials, generates blueprints, functions, and variables, adds components, and runs Python scripts. So the headline capability is real and demonstrated, and the documentation simultaneously tells you to stop expecting it to move. The readme also points at two alternatives: an official integration Epic is working on for engine version 5.8 and later, and a separate standalone implementation by another author under the MIT licence.

The engine range is narrower than the paid plugins claim

Version targeting is stated once for MCP and once for the paid catalogue, and the two do not agree in the direction you would expect. This plugin's MCP support targets engine versions 5.4 through 5.7 and later, and it is described as working alongside Claude Desktop, Claude Code, and Cursor. The paid plugins are advertised as covering 5.1 through 5.7 and later. So the free tier starts two minor versions later at the bottom of the range and ends at the same place, which means a project pinned to 5.1 through 5.3 can buy engine coverage but not the MCP server. Given that Epic's own integration is aimed at 5.8 and later, the interesting question for an adopter is not which range is listed but how long the current range stays relevant.

The free model list trails the paid one by a generation

Compare the two catalogues and the gap is easy to see. The free plugin lists OpenAI chat models in the 4.1, o4-mini, and o3 line plus structured outputs, with no image, audio, or realtime entries. The paid plugin lists the 5.x line including a pro and a codex variant, and adds the responses API, vision, realtime, image generation, text to speech, transcription, streaming, function calling, and multimodal. Anthropic is the same story: the free list has the 4-latest alias plus 3.7, 3.5 Sonnet and Haiku, while the paid list has 4.6 Opus and Sonnet and a 4.5 line. Grok shows the same offset, with 3-latest and a 3-mini beta in the free tier against 4.1, 4, a European variant, and a code-fast model in the paid one. DeepSeek is the exception worth naming: chat at version 3.1 and the R1 reasoning model appear only in the free list.

The description advertises 3D generation the free plugin does not offer

The repository description is a catalogue of everything the author's products do, and the free plugin is a subset of it. The description names text-to-3D and image-to-3D providers, two speech providers, and a set of image and Chinese-language providers alongside the chat models. None of the 3D or speech work appears in the free plugin's feature list. The free section is five lines: four hosted chat or reasoning providers plus a pointer for local models. Everything else is sold as a separate plugin on the marketplace, and the 3D line item is a plugin of its own with its own provider list covering text-to-3D, image-to-3D, remeshing, physically based textures, retexturing, and automatic rigging. Reading the description as a specification for this repository gives the wrong answer about what installs.

Local models are a pointer to another repository

Local inference is the one feature line in the free plugin that is not implemented in it. Instead of a bundled runtime, the local AI entry links to a separate repository, also under the MIT licence and from the same author's organisation, and names two model families it covers. The paid plugin then sells the same capability as a packaged product, listing a local runtime stack of several engines, the GGUF file format, and a local vision model, with the promise that no API keys are needed. So there are three positions on local models across the author's range: absent from the free plugin, delegated to a sibling repository, and packaged for sale. A team that has already standardised on a local runtime will find the middle option adequate; one that wants it inside the plugin build will not.

The last release is from February 2025 and the last commit from April 2026

The release history is two entries and both are alphas. The first is dated 2 January 2025 and its title carries a trailing space after the word alpha, which is the kind of detail that breaks a script parsing release names. The second is dated 15 February 2025 and is titled as the version that added DeepSeek R1. The last commit to the repository is dated 28 April 2026, roughly fourteen months after that tag. So there is a year of commits with no release, and the readme's promise that the plugin will keep getting updates with the latest features and models is being met in the default branch rather than in a tagged artefact. For a plugin whose whole value is staying current with model identifiers, that distinction is the one to keep in mind.

Cross-platform testing and Perforce support are the paid tier

The feature list for the free plugin is four items long, and two of them describe the boundary of what it does not do. Version control support covers Git submodules, with Perforce marked as work in progress rather than shipped. Lightweight means no external dependencies and, separately, that MCP can be excluded from the build, which is the one way to drop the unmaintained half of the plugin. The fourth item is the sharpest: testing across platforms and engine versions is listed as available in the paid plugins, which means the free plugin's own support across that matrix is unstated. For a runtime that calls hosted APIs and drives the editor, that is a reasonable trade for a free tier, and it is also the first thing to ask about when the plugin starts behaving differently on a build agent than on a workstation. The rest of the free feature list is short by design: a separate example project, submodule support, no external dependencies, and the option to leave MCP out of the build entirely.

Editorial conclusion

UnrealGenAISupport suits a team that wants hosted chat and reasoning models wired into an Unreal project this week without writing the integration layer, and that is content to keep the plugin's model identifiers updated by hand as vendors rename them. It does not suit a team that needs text-to-3D, speech, or an MCP server that will still be maintained, since none of those are in the free plugin and the MCP code is documented as unmaintained. Before adopting it, check the engine versions you target against the range it claims, and confirm that the absence of cross-platform testing is acceptable for your build matrix.

Frequently asked questions

What does UnrealGenAISupport do?

It is an Unreal Engine plugin that wraps hosted language model APIs so a project does not have to build its own integration layer. The free plugin covers chat for OpenAI, Anthropic, and Grok, DeepSeek chat and reasoning, and structured outputs, with local models delegated to a sibling repository.

Which Unreal Engine versions does UnrealGenAISupport support?

The MCP support targets engine versions 5.4 through 5.7 and later, while the paid plugins are advertised as covering 5.1 through 5.7 and later. Cross-platform and cross-version testing is listed as a paid-plugin feature rather than a free one.

Does UnrealGenAISupport generate 3D models or speech?

Not in the free plugin. Text-to-3D, image-to-3D, remeshing, textures, rigging, and text-to-speech are sold as separate paid plugins, even though the repository description lists those providers.

Is the UnrealGenAISupport MCP server maintained?

The readme states that MCP in this free plugin is not being actively developed. It points instead to an official integration Epic is working on for engine version 5.8 and later, and to a separate standalone implementation by another author.

How do I run local models with UnrealGenAISupport?

The free plugin does not bundle a local runtime. Its local AI entry links to a separate repository from the same author under the MIT licence. A paid plugin packages several local engines and the GGUF format with no API key required.

Official sources

  1. License: MIT
  2. prajwalshettydev/UnrealGenAISupport on GitHub
  3. Project website
  4. README
  5. Releases
Add this badge to your README

If you maintain this project, the badge below links readers to this analysis and shows its maintenance status from the daily GitHub snapshot. Paste the markdown into your README; add ?metric=license or ?metric=stars to the image URL for a different field.

Add this badge to your README

markdown
[![Hysen Labs](https://hysenlabs.com/badge/prajwalshettydev-unrealgenaisupport.svg)](https://hysenlabs.com/projects/prajwalshettydev-unrealgenaisupport)

Star history chart for your README

markdown
[![Star history](https://hysenlabs.com/chart/prajwalshettydev-unrealgenaisupport.svg)](https://hysenlabs.com/projects/prajwalshettydev-unrealgenaisupport)

Was this useful?