rasulkireev/osig
Generate Beautiful OG Images for Your Site
What's novel
Generate Beautiful OG Images for Your Site
Code Analysis
15 files read · 4 roundsRenders deterministic Open Graph and Twitter/X social preview images from typed canvas specs (rectangles, text, images) using Pillow, exposed via MCP tools, a Studio REST API, and a Django web UI, with Stripe billing, usage quotas, and Google Fonts provider support.
Strengths
The canvas rendering pipeline is genuinely well-built: a rich Pydantic-validated spec schema with discriminated unions, a Pillow-based renderer supporting gradients, shadows, rounded corners, text wrapping/alignment, image fitting, and opacity compositing, plus a thoughtful font provider system with safe Google Fonts fetching and caching. The MCP integration is clean, the URL safety validation (SSRF protection) is thorough, and the render observability layer with error classification and metrics
Weaknesses
The project is archived and the core use case (deterministic social image generation) is now largely solved by AI agents directly, reducing its practical value. The frontend is minimal (mostly Django templates with a small Stimulus controller), and the Stripe billing integration, while functional, is tightly coupled to dj-stripe with some awkward patterns like setting stripe.api_key at module level in views.py.
Score Breakdown
Signal breakdown
Innovation
Craft
Traction
Scope
Evidence
Commits
20
Contributors
3
Files
136
Active weeks
4
Repository
Language
Python
Stars
3
Forks
0
License
Apache-2.0