imageat

Generate AI images and videos with the imageat API. Use when the user asks to create, edit or animate images, remove a background, generate a video from text or a photo, or check their imageat balance

SKILL.md

GitHub

imageat

imageat generates images and videos from prompts and reference images. Every call is billed to the user's imageat account, so confirm before generating many outputs or long videos.

Setup

  • The user creates an API key at https://imageat.com/projects/keys (keys start with iat_live_). Read it from the IMAGEAT_API_KEY environment variable; never print it. If it is not set, ask the user to set it (or paste a key) before calling the API.
  • Base URL: https://imageat.com/api/v1
  • Send Authorization: Bearer $IMAGEAT_API_KEY and Content-Type: application/json.
  • Send an Idempotency-Key header (any unique string) on POSTs you might retry.
  • MCP clients can connect to https://mcp.imageat.com/mcp instead (OAuth or the same key).

Generate an image (synchronous)

curl -X POST "https://imageat.com/api/v1/images" \
  -H "Authorization: Bearer $IMAGEAT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"nano-banana-2","prompt":"A cinematic product photo on sculpted stone","aspectRatio":"1:1","resolution":"1K","numImages":1}'

The response contains url (one image) or urls (several), hosted on cdn.imageat.com.

Models (live list: GET https://imageat.com/api/v1/models/images):

  • nano-banana-pro (Nano Banana Pro): High-fidelity generation and editing with up to 4K output. From $0.40 / image.

  • nano-banana-2.1 (Nano Banana 2.1): Google's newest Nano Banana: up to 4K, extreme 8:1 to 1:8 ratios, up to 14 references, and optional web search grounding. From $0.16 / image.

  • krea-3 (Krea 3): Krea's newest model: aesthetic generation with prompt expansion, faithful instruction edits, up to 4 references, 1K or 2K. From $0.06 / image.

  • nano-banana-2 (Nano Banana 2): Fast, cost-efficient generation for iterative creative work. From $0.16 / image.

  • gpt-image-2.5 (GPT Image 2.5): OpenAI's newest image model: five quality tiers, scoped edits, and transparent cut-outs. Add gpt-image-2.5-sunburst for the slower, finer-detail variant. From $0.02 / image.

  • midjourney-v8 (Midjourney V8): Midjourney V8.2 with stylize, chaos, weird, and style references. Every run returns 4 images; native 2K HD optional. From $0.18 / 4 images.

  • gpt-image-2 (GPT Image 2): Precise instruction following with selectable quality tiers. From $0.02 / image.

  • seedream-5.0-pro (Seedream 5 Pro): High-fidelity ByteDance generation and reference-guided editing. From $0.14 / image.

  • nano-banana-2-lite (Nano Banana 2 Lite): The cheapest Nano Banana, for drafts, thumbnails and bulk runs. From $0.08 / image.

  • nano-banana (Nano Banana): The original Nano Banana (Gemini 2.5 Flash Image) for quick generation and edits. From $0.16 / image.

  • seedream-5.0-lite (Seedream 5 Lite): Faster, cheaper Seedream 5 for generation and reference-guided edits. From $0.07 / image.

  • krea-2-large (Krea 2 Large): Krea 2 text-to-image with up to 10 style references. One image per request. $0.13 / image.

  • krea-2-medium (Krea 2 Medium): Lighter Krea 2 with the same style references. One image per request. $0.07 / image.

  • GPT Image models accept quality; GPT Image 2.5 also takes xhigh/max and transparentBackground (needs png or webp output).

  • Nano Banana 2.1 takes up to 14 images, extreme ratios (8:1 to 1:8, plus auto) and enableWebSearch for grounding in live web results.

  • Krea 3 (krea-3) renders 1K or 2K (4K requests render at 2K) and takes up to 4 images; each reference is billed like extra output pixels, so send only the ones the edit needs.

  • Midjourney V8 (midjourney-v8) always returns 4 images per run (read urls; numImages is ignored) and takes ~2 minutes. resolution is 1K or 2K (native HD); optional stylize (0-1000), chaos (0-100), weird (0-3000), seed and styleReferenceUrl.

  • Editing / reference-guided generation: pass public URLs or base64 data URLs in images.

  • Other fields: aspectRatio, resolution (1K, 2K, 4K), outputFormat (png, jpeg, webp), numImages.

Edit an image

GET https://imageat.com/api/v1/edit/features lists the edit features (id, name, credit cost): background removal and replacement, object eraser, inpainting, relight, virtual try-on and more.

curl -X POST "https://imageat.com/api/v1/edit" \
  -H "Authorization: Bearer $IMAGEAT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"feature":"ideogram-remove-background","imageUrl":"https://example.com/photo.jpg"}'

Body: feature, imageUrl (public URL or base64 data URL), plus prompt, maskUrl, backgroundPrompt or backgroundImageUrl when the feature needs them.

Generate a video (asynchronous)

curl -X POST "https://imageat.com/api/v1/videos" \
  -H "Authorization: Bearer $IMAGEAT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"seedance25","providerVariant":"more","prompt":"A paper airplane glides through a softly lit studio","duration":"5s","resolution":"720p","aspectRatio":"adaptive"}'

The response has status: pending, taskId and generationId. Poll every 5-10 seconds:

curl "https://imageat.com/api/v1/videos/status?taskId=TASK_ID&generationId=GENERATION_ID" \
  -H "Authorization: Bearer $IMAGEAT_API_KEY"

Stop when status is completed (read url) or failed. Cancel with POST https://imageat.com/api/v1/videos/cancel and body {"taskId":"TASK_ID","generationId":"GENERATION_ID"}.

Models (live list: GET https://imageat.com/api/v1/models/videos):

  • wan3 (Wan 3.0): Text, image, first/last-frame or reference-image video; 2–30s · 480p/720p/1080p · audio; From $0.20.

  • minimax-h3-max (MiniMax H3 Max): Text, image, or first/last-frame video; 5–15s · 480p/768p; From $0.25.

  • seedance25 (Seedance 2.5): Text, image, reference, or frame-controlled video; 4–30s · 480p/720p; From $0.83.

  • midjourney-v8 (Midjourney V8): Image-to-video, 4 clips per run; 4 × 5s · 480p/720p; From $0.18.

  • pixverse (Pixverse V6): Text/image video with optional audio; 1–15s · up to 1080p; From $0.09/sec.

  • veo (Veo 3.1 Fast): Default model. Text or image video; 4/6/8s · 720p/1080p/4K · audio; From $1.20.

  • veo-lite (Veo 3.1 Lite): Text, image, or first/last-frame video; 4/6/8s · 720p/1080p · audio; From $0.40.

  • sora (Sora): Text or image video; 4/8/12s · 720p/1080p; From $0.80.

  • kling (Kling 2.6): Text or image video; 5/10s · 720p/1080p · optional audio; From $0.70.

  • kling3 (Kling 3.0): Text or image video; 3–15s · 720p/1080p/4K · optional audio; From $0.34/sec.

  • kling3-omni (Kling 3.0 Omni): Text or image video; 3–15s · 720p/1080p/4K · optional audio; From $0.42/sec.

  • grok (Grok Video): Text or image video; 1–15s · 480p/720p; From $0.10/sec.

  • grok-1-5 (Grok Video 1.5): Image-to-video; 1–15s · 480p/720p; From $0.17/sec.

  • seedance2 (Seedance 2.0): Text, image, or reference-image video; 4–15s · 480p/720p/1080p; From $1.08.

  • seedance2-mini (Seedance 2.0 Mini): Text, image, or reference-image video; 5/10/15s · 480p/720p; From $0.36.

  • gemini-omni (Gemini Omni Flash): Text or image video; 5/10s · 720p; From $2.12.

  • happyhorse (Happy Horse 1.0): Text or image video; 5/10/15s · 720p/1080p; From $1.40.

  • happyhorse-ref (Happy Horse 1.0 Reference): Reference-image video (referenceImageUrls); 5/10/15s · 720p/1080p; From $1.40.

  • flux3 (FLUX 3 Video): Text or image video; 5–20s · 720p/1080p; From $1.70.

  • Image-to-video: pass imageUrl (first frame), or firstFrameUrl + lastFrameUrl.

  • midjourney-v8 is image-to-video only: imageUrl is required, resolution is 480p or 720p, motion is low or high. Each run returns 4 five-second clips in urls.

Check balance

GET https://imageat.com/api/v1/credits returns the account's API balance in USD (and spendable web credits, which are spent first).

Errors

  • 401: missing or invalid key. Ask the user for a valid key.
  • 402: not enough balance (balance and required in the body). Send the user to https://imageat.com/api-pricing; do not retry.
  • 429: rate limited. Wait for Retry-After seconds, then retry.
  • 500: retry once with the same Idempotency-Key.

Reference