**Cost**: Model-priced per successful call. See the Credit Matrix for `image_upscale` model and resolution prices.

Image Upscale supports two input modes:

1. **JSON URL mode**: send `application/json` with the existing URL field.
2. **Direct upload mode**: send `multipart/form-data` with a local image file in the `image` field. The compatible field name `file` is also accepted for the primary image. Codia checks credits before reading and uploading the file, so insufficient-credit requests return `402` without uploading anything.

Direct upload example:

```bash
curl 'https://openapi.codia.ai/v2/open/image/upscale' \
  -H 'Authorization: Bearer {codia_api_key}' \
  -F 'image=@./image.png'
```

Enlarge an image to a higher resolution while reconstructing detail (not just bilinear interpolation). Model-driven — different models trade off sharpness, smoothness, and processing time. Synchronous.

## What this API is for

Recovering print-quality assets from low-resolution sources, preparing low-res user uploads for downstream tasks (e.g. background removal that needs more pixels), or salvaging legacy thumbnails. Use it before any pipeline step that benefits from more pixels — don't chain it after a generative step (those models already output at their target resolution).

## Request

| Field | Required | Notes |
|---|---|---|
| `image_url` | yes | Public URL of the image to upscale. |
| `model` | no | Model ID. Defaults to `codia_image_v2`. See the model price table for available models. |
| `target_size` | no | Target output dimensions, such as `3840x2160`. |

Use `target_size` to request output dimensions where supported by the model. For GPT Image 2 and 2.5, omitted or `auto` target size uses a 1K output adapted to the input aspect ratio and supported size constraints. Other models use their own defaults when target size is omitted.

For nano_banana_2 and nano_banana_pro upscale, explicit 1K/2K/4K values select that output and billing tier. For pixel dimensions, the closest supported aspect ratio is selected, then the lowest tier whose supported width and height cover the target, capped at 4K. For example, 1280x720 uses 1K, 2560x1440 uses 2K, and 3840x2160 uses 4K. Billing follows the selected tier. Omitted or auto target size uses 1K. Model size limits still apply.

## Response

`ImageUpscaleSuccess` shape — `data.image_urls` is an array of CDN URLs (length 1 for this endpoint). Use the first entry.

```json
{ "code": 0, "message": "ok", "data": { "image_urls": ["https://cdn.codia.ai/upscale/xyz.png"], "credits_charged": 16 } }
```

## Billing & limits

- Credits follow the `image_upscale` model/resolution table in the global Credit Matrix.
- The response includes `image_urls` and `credits_charged`, which reports the final charge.
- Use `POST /v2/open/estimate` with the same model and target size to estimate credits.
- Synchronous. Latency scales with input size; very large inputs may approach minute-scale.

## Errors

Standard set — `400` bad input, `401`/`403` auth, `402` credits, `429` rate limit, `500` internal.

> **Sync vs. async**
> This synchronous endpoint returns the result inline — the simplest path. For production pipelines that need webhook delivery, automatic retry with idempotency, or explicit task cancellation, call `POST /v2/open/tasks` with `operation=image_upscale` instead. Inputs and credit cost are identical.

Optional user-supplied key (up to 255 chars). When set, a successful (2xx) response is cached for 7 days. A retry with the same key + identical request body returns the cached response (no extra credit charged); a retry with the same key but a different body returns 409. Use to make POST safe to retry on network failures. For multipart/form-data uploads, requests are compared by form fields and file contents, so a retried upload is replayed even though the client picks a new multipart boundary.

Idempotency-Key

image_url

Model that performs the upscale. Defaults to `codia_image_v2`. Credits follow the `image_upscale` pricing table.

| Model | Default | 1K | 2K | 4K |
| --- | ---: | ---: | ---: | ---: |
| default / `codia_image_v2` | 16 | - | - | - |
| `ideogram_v3` | 16 | 16 | 16 | - |
| `nano_banana_2` | 18 | 18 | 27 | 41 |
| `nano_banana_pro` | 36 | 36 | 36 | 65 |
| `gpt_image` | 57 | 57 | 115 | 190 |
| `gpt_image_2_5` | 57 | 57 | 115 | 190 |

model

Optional target output dimensions as WIDTHxHEIGHT. Supported output sizes depend on the selected model. Pricing uses the requested output resolution tier. For GPT Image 2 and 2.5, omitted or auto target size uses a 1K output adapted to the input aspect ratio and supported size constraints. For nano_banana_2 and nano_banana_pro upscale, explicit 1K/2K/4K values select that output and billing tier. For pixel dimensions, the closest supported aspect ratio is selected, then the lowest tier whose supported width and height cover the target, capped at 4K. For example, 1280x720 uses 1K, 2560x1440 uses 2K, and 3840x2160 uses 4K. Billing follows the selected tier. Omitted or auto target size uses 1K. Model size limits still apply.

target_size

image

Compatible source image file field (deprecated: use `image` instead).

file

Source image URL when not uploading image

code

message

image_urls

credits_charged

data

The request body is malformed or missing required parameters

The API key is missing, invalid, or revoked

Account balance is insufficient or subscription has expired

Account does not have access to this endpoint

Idempotency-Key reused with a different request body.

Rate limited. Wait for the number of seconds in the `Retry-After` header before retrying.

An unexpected error occurred on the server

Enlarges an image and reconstructs detail instead of just stretching pixels.

Upscale Image

Input

Output

Related APIs