# Krea 3 | Image Edit Krea 3 transforms images from a text instruction, editing or combining up to 8 photos into one PNG for product, marketing and design teams on each::labs. ## API Information - **Model Slug:** krea-3-image-edit - **Branded URL:** https://www.eachlabs.ai/krea/krea-3/krea-3-image-edit - **Provider:** Krea Provider - **Category:** Image to Image - **Output Type:** image - **Status:** active - **Base Cost:** 1k: $0.03, plus $0.02 for each 1K reference image and $0.07 for each 2K reference image - **Interactive Demo:** https://www.eachlabs.ai/ai-models/krea-3-image-edit ## Pricing - **Charge Type:** dynamic - **Estimated Price (default example):** $0.0300 - **Pricing Details:** 1k: $0.03, plus $0.02 for each 1K reference image and $0.07 for each 2K reference image ### Pricing Rules | Condition | Pricing | | --- | --- | | resolution eq_i "1k" AND thinking_mode eq_i "high" | 1k, high thinking: $0.03 + $0.02, plus $0.02 for each 1K reference image and $0.07 for each 2K reference image | | resolution eq_i "1k" | 1k: $0.03, plus $0.02 for each 1K reference image and $0.07 for each 2K reference image | | thinking_mode eq_i "high" | 2k (default), high thinking: $0.06 + $0.02, plus $0.02 for each 1K reference image and $0.07 for each 2K reference image | | Rule 4 | 2k (default): $0.06, plus $0.02 for each 1K reference image and $0.07 for each 2K reference image | ## Input Schema | Parameter | Type | Required | Default | Constraints | Description | |-----------|------|----------|---------|-------------|-------------| | prompt | string | Yes | - | - | The edit instruction, or the image description when no input image is given. Say what must change and what must stay the same: the subject, the composition, the lighting and the background. Example: 'Replace the sky with a sunset and keep the house unchanged.' Refer to input images by their position, such as 'the first image'. | | image_urls | array | No | - | - | Up to 8 input image URLs (JPEG, PNG or WebP). Give 1 image to edit it. Give 2 or more images to combine them, for example a product and a scene. Leave empty to make a new image from the prompt only (text to image). | | resolution | string | No | 2k | 1k, 2k | Output resolution. 2k: native resolution, about 2048 px on the long side for the best detail (default). 1k: about 1024-1360 px on the long side, for previews and faster downloads. | | aspect_ratio | string | No | - | auto, 1:1, 4:5, 3:4, 2:3, 9:16, 5:4, 4:3, 3:2, 16:9, 21:9 | Auto: let Krea choose the framing automatically based on the input. When Auto is selected, aspect_ratio is omitted from the provider request. Output shape. Leave empty to let the model choose: an edit keeps the shape of the input image, and text to image gives a square image. 1:1 square for social posts. 4:5, 3:4, 2:3 and 9:16 portrait (9:16 for stories and reels). 5:4, 4:3, 3:2 and 16:9 landscape (16:9 for video frames and banners). 21:9 ultra-wide for cinematic frames. | | thinking_mode | string | No | fast | fast, high, ultra | How much the model plans before it edits. fast: the result in about 3-8 seconds, good for most edits (default). high: more planning for complex edits with many changes, about 10-30 seconds. ultra: the most planning for the hardest edits, slowest. ultra works only with resolution 2k. | | model | string | No | berry | berry, jam | Legacy compatibility option. Accepted for existing API clients, but no longer selects a provider variant. Krea uses its current Krea 3 model. | | seed | integer | No | - | - | Optional random seed (a non-negative integer). Use the same seed with the same inputs to get a similar result again. Leave empty for a random result. | ## Example Request ```bash curl -X POST https://api.eachlabs.ai/v1/prediction/ \ -H "Authorization: Bearer YOUR_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "krea-3-image-edit", "input": { "prompt": "Using the uploaded woman as the only identity reference (keep her exact face, freckles, blue-grey eyes, and long wavy honey-blonde hair), create a photorealistic Vogue-style fashion magazine cover, vertical 3:4: she reclines across the open petals of a giant pale blush-pink lily on a thick green stem against a flat cobalt-blue sky, propped on one elbow, legs stretched out with ankles crossed, looking coolly at the camera, wearing her own pale yellow floral slip dress extended to midi length and cream T-strap heels, with tall green stamens and burnt-orange anthers rising behind her, oversized graphic tropical leaves at the bottom, low angle, hard midday sun, surreal 1970s fashion campaign look; across the top, a huge masthead that reads exactly \"each::labs\" in a bold lowercase modern sans-serif, the letters \"each\" and \"labs\" in white, and the \"::\" in the middle drawn as four small solid orange-red (#FD441F) squares in a 2x2 grid, with the lily stamens slightly overlapping the masthead like a classic Vogue cover; no other text anywhere, no barcode, no watermark." } }' ``` ## Output Schema Response returned by `GET /v1/prediction/{id}` when the job completes: ```json { "status": "success", "predictionID": "string", "output": "string (URL of generated image)", "metrics": { "predict_time": "number (seconds)" } } ``` ## Polling ```bash curl https://api.eachlabs.ai/v1/prediction/{PREDICTION_ID} \ -H "Authorization: Bearer YOUR_API_KEY" ``` | Status | Meaning | |--------|---------| | `processing` | Still running — poll again | | `success` | Done — read `output` | | `error` | Failed — read `message` / `details` | ## Webhook (alternative to polling) Pass `"webhook_url": "https://your.host/path"` in the create request. Eachlabs POSTs this payload when the job ends: ```json { "exec_id": "prediction-uuid", "status": "succeeded", "output": "https://...", "error": "" } ``` `status` is `"succeeded"` or `"failed"`. `exec_id` equals the `predictionID` from create. Return 2xx within 30 seconds. ## Errors Error body: `{ "status": "error", "message": "...", "details": "..." }` | Code | Meaning | |------|---------| | `400` | Invalid input | | `401` | Missing / invalid `Authorization` bearer token | | `404` | Unknown model or prediction id | | `429` | Rate limit — 100 creates / min, 10 concurrent per key | | `5xx` | Retry with backoff | ## Overview **Krea 3 Image Edit Overview** Krea 3 transforms images from a text instruction, editing or combining up to 8 photos into one PNG for product, marketing and design teams on each::labs. The model supports prompt-only generation or editing and combining up to eight input images. Krea 3 Image Edit from Krea is available through each::labs in the Krea 3 family. Required fields are prompt. image_urls: Up to 8 input image URLs (JPEG, PNG or WebP). resolution: Output resolution. Use the Krea 3 Image Edit API when you need this transformation or analysis as a repeatable step in an application or creative pipeline. The sections below distinguish the required material, adjustable controls and checks that matter before delivery. ## Usage Notes - API Base URL: `https://api.eachlabs.ai/v1` - Authentication: send `Authorization: Bearer YOUR_API_KEY`. Generate a key from the Eachlabs dashboard at https://www.eachlabs.ai/dashboard/api-keys. - File-typed parameters (`*_url`, `image_url`, `video_url`, `audio_url`, etc.) accept publicly-reachable HTTPS URLs only. Upload your asset first (GCS / S3 / your CDN) and pass the resulting URL. Data-URIs and localhost URLs are rejected. - For structured parameters (arrays / objects) send real JSON values, not stringified payloads. - Monetary values are reported in USD; per-token / per-megapixel rates may be billed in micro-cents internally. - Prefer `webhook_url` over polling for long-running predictions — see the Webhook Callback section.