Frame Ronin
FreeNot checkedGenerate pixel art sprites, sprite sheets, and game assets using AI image models via MCP.
About
Generate pixel art sprites, sprite sheets, and game assets using AI image models via MCP.
README
FrameRonin MCP
22 MCP tools. One server. Complete pixel art game asset pipeline for AI.
AI generates → AI processes → Game-ready assets. All through a single MCP connection.
English | 中文
Overview
FrameRonin MCP wraps the FrameRonin pixel art toolset into an MCP (Model Context Protocol) server. AI assistants can directly generate sprites via multiple backends, remove backgrounds, compose sprite sheets, convert GIFs, pixelate images, and extract video frames — all through tool calls.
Original project: systemchester/FrameRonin — a full-stack web application with UI for pixel art processing. This MCP server extracts the core algorithms and adds AI-powered generation, making them callable by Claude, Gemini CLI, Copilot, and other MCP-compatible AI tools.
What Can AI Do With This?
"Generate a pixel art fire mage sprite sheet and prepare it for RPG Maker"
→ generate_rpgmaker("fire mage, red robes", rows=4, cols=4, width=48, height=48)
→ 16 frame PNGs, transparent background, no watermark, ready to import
"Remove the green screen from this actor photo"
→ image_chroma_key("actor_green_screen.png", key_color="green", tolerance=40)
"Convert this video of a walk cycle into a sprite sheet"
→ video_to_spritesheet("walk_cycle.mp4", fps=12, columns=4)
"Turn this high-res artwork back into pixel art"
→ image_pixelate("artwork.png", num_colors=16, scale_result=4)
Prompt Auto-Expansion
All generate_* tools automatically expand short prompts to 2000+ characters with professional pixel art specifications — color palette rules, anti-aliasing bans, dithering patterns, sprite sheet layout conventions, and RPG Maker format requirements. Just describe your character/item/monster in plain language; the framework handles the technical details.
Demo — Full Pipeline
Input: "a warrior with silver armor and a sword" — automatically expanded, then processed through all 5 steps:
All 16 frames extracted at native resolution — 4 directions × 4 animation frames:
Row 0 (Down) animation loops — native cells padded square, NEAREST, no quality loss
| Step | Tool | Input → Output |
|---|---|---|
| 1. Generate | generate_gemini |
Auto-expanded prompt → sprite sheet |
| 2. Clean | image_remove_gemini_watermark |
Remove AI watermark via reverse alpha blending |
| 3. Key Out | white-to-alpha | White pixels → transparent (pixel-safe, no AI blur) |
| 4. Resize | image_resize (NEAREST) |
Pixel-perfect scale preserving hard edges |
| 5. Split | spritesheet_split |
Sheet → individual frame PNGs |
Final output: 16 PNGs (48×48, transparent background) — directly importable into RPG Maker MV, Godot, Unity, or any 2D game engine.
Install
pip install frame-ronin-mcp
Requires Python 3.11+, FFmpeg (for video tools), Chrome (for Gemini web generation).
Configure
Add to your MCP client config (e.g. .claude/mcp.json for Claude Code):
{
"mcpServers": {
"frame-ronin": {
"command": "frame-ronin-mcp"
}
}
}
No environment variables needed for free Gemini generation or any post-processing tools.
Tools (22)
Generation — 6 tools
Five independent backends + one unified pipeline. Each backend is a self-contained module — add new ones by dropping a .py file.
| Tool | Backend | Auth | Best for |
|---|---|---|---|
generate_gemini |
Gemini free web app | Google login (once) | Free, no setup |
generate_dalle |
OpenAI DALL-E 3 | OPENAI_API_KEY |
Natural language, diverse styles |
generate_stability |
Stability AI | STABILITY_API_KEY |
Pixel art control, fine-tuning |
generate_gemini_api |
Gemini API | GOOGLE_API_KEY + billing |
Fastest, programmatic |
generate_siliconflow |
SiliconFlow | SILICONFLOW_API_KEY |
Qwen/FLUX models, Chinese prompts |
generate_rpgmaker |
Any backend → full pipeline | per backend | One-call: gen → clean → matte → split |
generate_rpgmakerpipeline: generate → remove Gemini watermark → AI background removal (rembg/U2Net) → resize to RPG Maker grid → split into individual frame PNGs. All in one call.
Processing — 16 tools
| Category | Tool | Description |
|---|---|---|
| Video | video_to_spritesheet |
Extract frames → rembg matting → sprite sheet + index JSON |
video_get_info |
Probe video: duration, resolution, fps, frame count | |
video_remove_watermark |
Remove Seedance/Jiemeng watermark (crop bottom 180px) | |
| Matting | image_remove_background |
AI matting via rembg/U2Net |
image_chroma_key |
Green/blue screen removal + spill suppression + edge erosion | |
image_double_background_matte |
Black+white differential alpha extraction | |
| Image | image_remove_gemini_watermark |
Reverse alpha blending with embedded 48/96px mask |
image_resize |
Scale (lanczos/nearest/bilinear/bicubic) | |
image_crop |
Crop to rectangle | |
image_merge_grid |
Merge images into grid atlas | |
| GIF/Sprite | gif_extract_frames |
Extract frames from animated GIF |
frames_to_gif |
Combine frames into GIF | |
spritesheet_split |
Split spritesheet by uniform grid | |
spritesheet_compose |
Compose frames into spritesheet + index | |
| Pixel Art | image_pixelate |
Proper pixel art: OpenCV mesh detection → downsample |
image_pixelate_simple |
Simple uniform-grid pixelation (fast) |
Project Structure
FrameRonin-MCP/
├── frame_ronin_mcp/
│ ├── server.py # MCP server entry point
│ ├── tools/
│ │ ├── generate/ # Multi-backend image generation
│ │ │ ├── gemini_web.py # Gemini free web app (Playwright)
│ │ │ ├── gemini_api.py # Gemini API (google-genai)
│ │ │ ├── dalle.py # OpenAI DALL-E 3
│ │ │ ├── stability.py # Stability AI
│ │ │ └── siliconflow.py # SiliconFlow (Qwen/FLUX)
│ │ ├── video.py # FFmpeg + rembg pipeline
│ │ ├── matting.py # AI/chroma/double-bg matting
│ │ ├── image.py # Resize, crop, watermark removal
│ │ ├── gif_sprite.py # GIF extraction, sprite sheet ops
│ │ └── pixelate.py # Proper pixel art conversion
│ └── lib/ # Core algorithms (pure Python)
│ ├── watermark.py # Gemini watermark removal
│ ├── chroma_key.py # Chroma key matting
│ ├── double_bg.py # Double background matting
│ ├── mesh.py # OpenCV grid detection
│ └── pixelate_core.py # Proper pixel art algorithm
├── pyproject.toml
└── README.md
Related
- FrameRonin — original web application with full UI
- proper-pixel-art — pixel art downsampling algorithm
- GeminiWatermarkTool — watermark removal technique
License
MIT © FrameRonin
Installing Frame Ronin
This server has no published package — it is built from source. Open the repository and follow its README.
▸ github.com/gooddayday/frameronin-mcpFAQ
Is Frame Ronin MCP free?
Yes, Frame Ronin MCP is free — one-click install via Unyly at no cost.
Does Frame Ronin need an API key?
No, Frame Ronin runs without API keys or environment variables.
Is Frame Ronin hosted or self-hosted?
Self-hosted: the server runs locally on your machine via the install command above.
How do I install Frame Ronin in Claude Desktop, Claude Code or Cursor?
Open Frame Ronin on unyly.org, pick your client tab (Claude Desktop, Claude Code, Cursor) and press Install — the config is generated automatically, no JSON editing.
Related MCPs
ARA
Generate images, video and audio from any AI agent — one connector.
by ARAOmni Video
An MCP server that transforms LLM-enabled IDEs into professional video editors by pre-processing footage into text proxies, generating motion graphics via HTML/
by buildwithtazaYouTube
Transcripts, channel stats, search
by YouTubeEverArt
AI image generation using various models.
by modelcontextprotocolCompare Frame Ronin with
Not sure what to pick?
Find your stack in 60 seconds
Author?
Embed badge for your README
Browse similar
All media MCPs
