Which sizes does the changelog template support?

1200×630, 1080×1080. The default is 1200×630; pass width and height to pick another supported size.

Which parameters are required?

version and items. Every other parameter is optional and has a rendering default.

Does rendering the same parameters twice cost twice?

No. The same template version, parameters, brand kit version, format, and size produce the same cache key, so a repeat request is served from cache instead of rendering again. Billing is not connected yet in early access.

Template · changelog@1

Changelog template

The changelog template renders a release card with a version, date, tag, and up to five highlights. Send JSON parameters through the remote MCP server, POST /v1/renders, or a signed /i/changelog.png URL; the default size is 1200×630.

Sizes: 1200×630 · 1080×1080
Formats: PNG · JPEG · WebP
Version: 1.0.0

VectorFast Render is in early access. Billing is not connected yet, and API keys are issued to waitlist members as access opens. Join the waitlist

Preview

Rendered by VectorFast Render from this template's English and Chinese example parameters, through the same signed URLs your own pages would use.

Preview images are signed on the server and are not available in this environment.

Ask your agent

Paste the prompt into Claude, Cursor, or another agent connected to the VectorFast Render MCP server. The JSON is the render tool call it should turn into.

Prompt

Make a changelog card for our v2.3 release today, tagged "Release", with three highlights: brand kits with versioning, logo vectorization on upload, and 40% faster cold renders.

MCP render call

{
  "name": "render",
  "arguments": {
    "template": "changelog",
    "params": {
      "version": "v2.3",
      "date": "Oct 30, 2026",
      "tag": "Release",
      "items": [
        "Brand kits with versioning",
        "Logo vectorization on upload",
        "40% faster cold renders"
      ]
    }
  }
}

Connect Claude to VectorFast Render

Parameters

Read from the schema that POST /v1/renders validates. String values are trimmed; anything over a limit is rejected with invalid_params.

Changelog template parameters
Param	Type	Required	Limit	Notes
version	string	Yes	≤ 24 chars	Version label such as "v2.3".
date	string	No	≤ 30 chars	Free text release date.
items	string[]	Yes	1–5 items, ≤ 80 chars each	One to five highlights. Pass an array of strings; signed URLs carry it as a JSON-encoded array string.
tag	string	No	≤ 24 chars	Small label such as "Release" or "Hotfix".

Render it from code

The REST call needs an API key in Authorization: Bearer. Its response includes a signed url you can store and reuse as an og:image. The Node.js function shows the signing algorithm the server verifies; per-key signing secrets are not issued yet in early access, and the secret must stay on your server.

curl · POST /v1/renders

curl https://vectorfast.ai/v1/renders \
  -H "Authorization: Bearer $VECTORFAST_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "template": "changelog@1",
  "format": "png",
  "width": 1200,
  "height": 630,
  "params": {
    "version": "v2.3",
    "date": "Oct 30, 2026",
    "tag": "Release",
    "items": [
      "Brand kits with versioning",
      "Logo vectorization on upload",
      "40% faster cold renders"
    ]
  }
}'

Node.js · signed URL (server-side only)

import { createHmac } from 'node:crypto'

// Server-side only. Never ship the signing secret to a browser, a mobile app or a repo.
const SIGNING_SECRET = process.env.VECTORFAST_SIGNING_SECRET

function encode(value) {
  return encodeURIComponent(value).replace(/[!'()*]/g, c => '%' + c.charCodeAt(0).toString(16).toUpperCase())
}

export function signRenderUrl({ template, format = 'png', params, brand, width, height, keyId }, secret = SIGNING_SECRET) {
  const query = {}
  for (const [key, value] of Object.entries(params)) {
    if (value === undefined || value === null)
      continue
    query[key] = typeof value === 'string' ? value : Array.isArray(value) ? JSON.stringify(value) : String(value)
  }
  if (brand) query.brand = brand
  if (keyId) query.k = keyId
  if (width) query.w = String(width)
  if (height) query.h = String(height)

  const canonical = Object.entries(query)
    .map(([key, value]) => [encode(key), encode(value)])
    .sort((a, b) => (a[0] < b[0] ? -1 : a[0] > b[0] ? 1 : a[1] < b[1] ? -1 : a[1] > b[1] ? 1 : 0))
    .map(([key, value]) => key + '=' + value)
    .join('&')
  const path = '/i/' + template + '.' + format
  const sig = createHmac('sha256', secret).update(path + '?' + canonical).digest('base64url')
  return 'https://vectorfast.ai' + path + '?' + (canonical ? canonical + '&' : '') + 'sig=' + sig
}

const url = signRenderUrl({
  template: 'changelog',
  format: 'png',
  width: 1200,
  height: 630,
  params: {
    "version": "v2.4.0",
    "date": "Oct 30, 2026",
    "tag": "Release",
    "items": "[\"Remote MCP server for Claude, Cursor and Codex\",\"Signed image URLs that never expire\",\"WebP and JPEG output next to PNG\"]"
  },
})

How signed URLs work

Changelog template FAQ

Which sizes does the changelog template support?: 1200×630, 1080×1080. The default is 1200×630; pass width and height to pick another supported size.
Which parameters are required?: version and items. Every other parameter is optional and has a rendering default.
Does rendering the same parameters twice cost twice?: No. The same template version, parameters, brand kit version, format, and size produce the same cache key, so a repeat request is served from cache instead of rendering again. Billing is not connected yet in early access.

Updated 2026-10-30 · Markdown version