AVGE Engine
Vector illustration tools for AI agents — create scenes, characters, and graphics with shapes, armatures, materials, and rendering.
README
AVGE Engine — AI-Native Vector Graphics Engine
Version: 0.5.0 | Tool set: m0b-v1
Vector illustration tools for AI agents — create editable vector art with shapes, armatures, materials, and rendering. Output is clean SVG you can open in Illustrator, Figma, or any vector editor. Runs as an MCP server.
View examples — fridge scene, bedroom, landscape, iPhone mockup, cat manga, and more with previews.
Quick Start
python -m avge_engine api # FastAPI HTTP server (:8000)
python -m avge_engine mcp-sse # MCP SSE server (:8001)
python -m avge_engine mcp # MCP stdio server
Preview renders at http://localhost:8000/preview/<document_id>.png
Features
- Normalized coordinates — 0.0–1.0 canvas, no pixel math
- Procedural geometry — radial spread, offset outlines, guide lines, building chains, speech bubbles, bursts, armature skeletons, foreshortening
- Armature skeletons — node-edge graphs with tapered segments, junction separation, filleted V-gaps, curved Catmull-Rom chains, and unary union merging
- Boolean operations — union, intersect, subtract, xor with Ramer-Douglas-Peucker simplification
- HSL shading — auto highlight + shadow from light direction
- Primitives — rects (including tapered/trapezoid), ellipses, lines, polylines, arcs, polygons, stars
- Text + images — SVG text with font family/style/anchor, embedded images via
<image> - Gradient backgrounds — linear and radial gradient definitions
- Style system — fill, stroke, stroke-width, opacity, blend modes, dash patterns
- Batch operations — execute multiple tools in one call, edit multiple regions with per-region transforms
- Palette generation — HSL harmony presets (complementary, triadic, analogous, etc.)
- Cross-document copy — copy elements between documents with offset
- SVG arc support — A/a commands in
import_svg_path(elliptical arcs with sweep/large-arc flags) - Relative positioning — place elements relative to a parent region's bounding box
- PNG rasterization — via rsvg-convert (librsvg) with Unicode/emoji font support
- Per-document tracking — checkpoint/restore history, tool usage stats
42+ MCP Tools
| Category | Tools |
|---|---|
| Document | create_document, list_documents, set_background, get_document_stats |
| Create | create_region, create_primitive, create_curve, create_text, insert_image, import_svg_path |
| Edit | edit_region (point-level nudge), edit_regions (batch transforms), delete_region, copy_element |
| Transform | transform_objects, duplicate, boolean_operation, mirror_region |
| Style | restyle, add_shading, generate_palette, define_gradient, apply_line_hierarchy, compare_style_consistency |
| Groups | group_regions, ungroup_regions |
| Procedural | generate_shape (14+ patterns: armature, segmented_chain, radial_spread, speech_bubble, create_burst, foreshorten, etc.) |
| View | render_preview, describe_scene, checkpoint_diff, render_diff |
| History | checkpoint, restore, get_history, batch |
Coordinates
All coordinates are normalized 0.0–1.0 where (0, 0) = top-left and (1, 1) = bottom-right.
Relative Positioning
Create tools accept relative_to — pass a region ID and coordinates become 0–1 fractions of that region's bounding box:
{"tool": "create_primitive", "shape": {"type":"ellipse","cx":0.5,"cy":0.5,"rx":0.1}, "relative_to": "belt_panel"}
Architecture
avge_engine/
├── controllers/ # MCP tool definitions (region, scene_ops, style, procedural, ...)
├── scene/ # Scene graph (graph.py, models.py)
├── geometry/ # Curve fitting, procedural patterns, types
├── effects/ # Style dataclass, HSL color transforms
├── renderer/ # SVG serializer + PNG rasterization
├── services/ # Engine service, document persistence
├── api.py # FastAPI HTTP server
└── server.py # MCP server setup (FastMCP)
- Python 3.12 with FastMCP (MCP SDK)
- Normalized coordinates, resolved at render time
- Catmull-Rom → cubic Bézier curve fitting (deterministic, closed-form)
- SVG output with text, images, gradient defs
- PNG preview via rsvg-convert with fontconfig/Unicode support
- JSON file persistence per document
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.