rendry

Rendry documentation

REST API and MCP server for rendering images and PDF documents from templates. External AI clients can supply layouts and copy through MCP; the rendering pipeline does not run a generative image model. Separate ML tools process existing images: rembg/U2Net removes backgrounds, and optional SAM3 through fal.ai segments objects. This page summarizes the interactive documentation.

Getting started

Create an account, design a template in the visual editor (or clone one of 51 gallery templates), generate an API key in Settings → API Keys, and render:

curl -X POST https://rendry.io/api/v1/render/sync \
  -H "Authorization: Bearer sk_live_..." \
  -H "Content-Type: application/json" \
  -d '{"preset_id": "your-template", "dynamic_data": {"title": "Hello"}}'

The response contains result_url of the finished PNG. Typical render: ~1-3 s.

What the docs cover

• Authentication; API keys (Bearer sk_live_...) and OAuth 2.1 for MCP clients
• Templates and presets; JSON DSL, layers (text, image, shape, QR, barcode, group), {variables} and dynamic_data
• Rendering; sync (up to 25 s), async with webhooks, batches up to 10 images per call
• Connecting AI agents; remote MCP endpoint https://rendry.io/api/v1/mcp, 33 tools, drafts at 0.1 of a credit
• Assets; uploading images and custom fonts, background removal, stock photo search
• Limits; canvas up to 6144×6144 px, 40 layers, plans and quotas

Machine-readable: llms.txt, llms-full.txt, OpenAPI, MCP manifest.