CONNECT YOUR
AGENT IN
5 MINUTES.

The fastest path. Describe what you want and MemeForge auto-selects the best template and renders the meme in one call. Great when you know the vibe but not the exact template.

generate_meme({
  description: "drake approval meme",
  texts: [
    "Spending 6 hours debugging",
    "Adding console.log everywhere"
  ]
})
// → Returns rendered meme image

Search templates WITHOUT rendering. Returns zone labels, descriptions, and hints so you can verify the match before committing. Use this when you want to preview which template fits.

find_template({
  query: "plan that backfires",
  limit: 3
})
// → Returns:
// 1. gru-plan — Gru's Plan [featured]
//    Zones: [0] step-1, [1] step-2, [2] step-3, [3] realization
//    Use when: multi-step plan where step 3 or 4 backfires

Generate a meme from a specific template ID. Texts are auto-positioned into the template's zones in order. Supports fuzzy matching if the exact ID isn't found.

meme_from_template({
  template: "expanding-brain",
  texts: [
    "Using var",
    "Using let",
    "Using const",
    "Using TypeScript enums"
  ]
})

Full control rendering. Provide a custom image URL with manually positioned text overlays and optional deep-fry effects. Use this for non-template memes or custom layouts.

render_meme({
  image: "https://example.com/photo.jpg",
  width: 800,
  height: 600,
  texts: [{
    text: "TOP TEXT",
    x: 50, y: 20,
    width: 700, height: 100,
    fontSize: 64,
    color: "#FFFFFF",
    outlineColor: "#000000",
    outlineWidth: 4,
    font: "Impact",
    allCaps: true,
    textAlign: "center",
    textStyle: "outline"
  }],
  deepFry: {
    noise: 3,
    jpegIterations: 2,
    jpegQuality: 15
  }
})

Browse all 220+ templates with optional search and tier filter. Returns template IDs, zone counts, and keywords. Paginated.

list_templates({
  query: "comparison",
  limit: 5,
  tier: "featured"
})
// → drake, distracted-boyfriend, two-buttons, ...

Resize to target dimensions. Fit modes: cover (crop to fill), contain (fit within, may letterbox), stretch, inside (shrink only), outside (grow only).

resize({
  image: "https://example.com/photo.jpg",
  width: 400,
  height: 400,
  fit: "cover"
})

Crop a rectangular region by pixel coordinates.

crop({
  image: "https://example.com/photo.jpg",
  x: 100, y: 50,
  width: 600, height: 400
})

Rotate by any angle. Canvas expands to fit. Positive = clockwise.

rotate({
  image: "https://example.com/photo.jpg",
  angle: 45,
  background: "#000000"
})

Mirror an image horizontally, vertically, or both.

flip({
  image: "https://example.com/photo.jpg",
  direction: "horizontal"
})

Adjust brightness, contrast, saturation, blur, opacity. Toggle grayscale, sepia, or invert.

adjust({
  image: "https://example.com/photo.jpg",
  brightness: 20,
  contrast: 30,
  saturation: -50,
  grayscale: false,
  blur: 2
})

Layer multiple images on a canvas. Each layer has position, size, opacity, rotation, and anchor point. Layers draw in array order (first = bottom).

compose({
  width: 1200,
  height: 800,
  background: "#000000",
  layers: [
    { image: "https://.../background.jpg" },
    {
      image: "https://.../logo.png",
      x: 50, y: 50,
      width: 200,
      opacity: 0.8,
      anchor: "top-left"
    },
    {
      image: "https://.../sticker.png",
      x: 900, y: 600,
      rotation: -15,
      anchor: "center"
    }
  ]
})

Arrange up to 20 images in a grid, horizontal strip, or vertical strip. Auto-sizes cells. Configurable gap and background color.

collage({
  images: [
    "https://.../photo1.jpg",
    "https://.../photo2.jpg",
    "https://.../photo3.jpg",
    "https://.../photo4.jpg"
  ],
  layout: "grid",
  columns: 2,
  gap: 8,
  background: "#FFFFFF"
})

Stamp a watermark/logo onto an image. 5 positions, configurable opacity, scale, and margin.

watermark({
  image: "https://.../photo.jpg",
  watermark: "https://.../logo.png",
  position: "bottom-right",
  opacity: 0.3,
  scale: 0.15,
  margin: 20
})

Strip text from a meme. First tries template matching (returns the clean original if found). Falls back to pixel-based inpainting. Adjustable sensitivity and radius.

remove_text({
  image: "https://.../meme-with-text.jpg"
})
// → "Text removed using template matching
//    (clean original found!)"

// Force pixel inpainting:
remove_text({
  image: "https://.../custom-meme.jpg",
  forceInpaint: true,
  sensitivity: 60,
  radius: 10
})

5 methods: auto (ML first, fallback to color), ml (rembg U2-Net), sam (MobileSAM with bounding boxes — best quality), color (flood fill), threshold (brightness cutoff). Returns transparent PNG.

// Automatic (recommended)
remove_background({
  image: "https://.../product.jpg",
  method: "auto"
})

// SAM with bounding box (best for specific subjects)
remove_background({
  image: "https://.../group-photo.jpg",
  method: "sam",
  boxes: [
    { x1: 100, y1: 50, x2: 400, y2: 500 }
  ]
})

// Color-based (e.g. green screen)
remove_background({
  image: "https://.../greenscreen.jpg",
  method: "color",
  color: "#00FF00",
  tolerance: 30
})