Medal

General Intuition is an AI research lab and company developing models that perceive, predict, and act in virtual and physical environments. Its action models process visual frames and continuously output actions, rather than responding turn by turn like language models; the company says these models are trained using action-labeled gameplay data from Medal and can support game-playing and control of multiple robot types through shared pixels-in, actions-out primitives. It also develops world models that predict the outcomes of actions, including MIRA, a real-time multiplayer world model trained on Rocket League data collected with publicly available bots. General Intuition has onboarded a small number of partners in games, simulation, and robotics to a commercial API and is collecting waitlist applications for broader access.

h3-metal is a native plain-C and Metal inference engine for the MiniMax H3 video model on Apple Silicon Macs. It provides a command-line interface and Iris-style interactive session for prompt-to-video and prompt-to-audio generation, including synchronized stereo audio, first- and last-frame conditioning, and ordered Ref2VA image, video, and audio references. The CLI can inspect model layout and the selected Metal device, run prompt-to-media jobs, and retain exact BF16 prompt conditioning, the prepared diffusion transformer, and video decoder in memory so repeated prompts with different seeds avoid reloading and re-encoding them. Configurable controls include denoising passes, transition reuse, transformer layers, frame dimensions, seeds, and output files. The project is developed through incremental vertical slices covering deterministic host and model metadata, portable Metal block parity, prompt encoding, media generation, conditioning, and H3-specific Metal performance and memory optimization.

AI foundation models from General Intuition for acting in virtual and physical environments. They process visual frames and continuously produce actions rather than responding turn by turn like language models, enabling control of games and more than 10 types of robots through existing game-controller interfaces. General Intuition trains its models on action-labeled video data, including gameplay clips from Medal, and offers access through a commercial API to selected partners while broader access is provided through a waitlist.

Video to Game is an open-source skill for coding agents that turns a short gameplay video into a playable Three.js browser game. The agent studies the video frame by frame, extracts a visual specification, generates matching images, 3D models, sound effects, and music through fal, writes the game as a standalone HTML file, and then playtests and records it with keyboard scenarios. It is distributed for Claude Code, Codex, Cursor, Gemini CLI, Copilot, and other agents through skill installation or copying. The workflow requires Python 3.9+, ffmpeg, Node 18+, Playwright Core, and a fal API key; the repository is MIT-licensed.

Universal Modder is an open-source toolkit by rehan-remade for AI coding agents that create offline, single-player mods for PC games. Its skills and tools identify a game's engine and existing modding route, inspect code and files, back up saves, build a working mod slice, generate art, 3D assets and audio through the fal MCP server, test the result in the running game, package it, and record findings for later agents. The repository supports Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot, OpenCode and other agents that read AGENTS.md, and includes the Python `um` CLI for game scanning, asset generation and processing, Windows automation, save management, video capture, publishing checks and a shared knowledge base of game-specific field notes. Its rules restrict use to games owned by the user and offline or single-player work; they prohibit online cheats, anti-cheat or DRM bypasses, and distributing game files or decompiled code. The repository is MIT licensed and requires Python 3.10+ and ffmpeg; Blender is used for 3D-to-sprite rendering.