# Krea 2 Trainer API

> Addestra una LoRA personalizzata Krea 2 da un archivio zip di immagini, con didascalie automatiche facoltative

Esegue il fine-tuning di una LoRA su un .zip di immagini in `images_data_url`, con didascalie .txt opzionali aventi lo stesso nome e una `trigger_phrase`. Le modalità `auto_captioning` sono Off, Object/Character, Style e Custom. `steps` va da 50 a 10000 (predefinito 100) e `resolution` è 768 o 1024. Restituisce i pesi LoRA e un file di configurazione.

- Model id: `krea-2-trainer`
- Made by: Krea
- Category: Training
- Output: file
- Price per run: from $0.30 / run (cheapest run $0.30 (300 API credits))
- Run: `POST https://api.doitong.com/v1/run/krea-2-trainer`
- Quote: `POST https://api.doitong.com/v1/quote/krea-2-trainer`
- Docs page: https://doitong.com/it/developers/models/krea-2-trainer

## Pricing

Price: **from $0.30 / run** (the lowest tier; the quote prices your exact input). Price per run: the input files are measured before the run (length, resolution, frame rate). The quote measures the same files and is exactly what the run charges.

The full example input below costs $0.30 = 300 API credits (1 run × $0.30).

Prices are the upstream price in USD, paid from the API balance in API credits (1000 API credits = 1 USD). `POST https://api.doitong.com/v1/quote/krea-2-trainer` with the same JSON body returns the exact price for that input; the quote is exactly what a run charges. Invalid input is never charged and failed or canceled requests are refunded automatically.

## Input parameters

| Parameter | Canonical name | Type | Required | Default | Allowed values / range | Description |
|---|---|---|---|---|---|---|
| `images_data_url` |  | string | yes |  |  | URL di uno .zip di immagini di addestramento contenente file PNG/JPG/JPEG/WebP; una richiesta fallisce con 422 se un'immagine non ha né un file di didascalia né un `trigger_phrase` non vuoto. Le didascalie sono facoltative quando `trigger_phrase` fornisce il fallback; in caso contrario usa gli stem dei nomi file corrispondenti: ROOT.EXT corrisponde a ROOT.txt, ad esempio 001.jpg con 001.txt. |
| `trigger_phrase` |  | string | no | `` |  | Frase di istanza in stile DreamBooth obbligatoria se `auto_captioning` è Off e un'immagine non contiene un file di didascalia. Fornisce le didascalie dove mancano i file .txt e il testo iniziale per i prompt di convalida predefiniti. |
| `auto_captioning` |  | enum (string) | no | `Off` | `Off`, `Object/Character`, `Style`, `Custom` | Modalità di captioning opzionale: Object/Character, Style e Custom richiamano dataset-captioner con google/gemma-4-31B-it, mentre Off mantiene il requisito esistente che ogni immagine abbia una didascalia .txt con stem corrispondente o usi trigger_phrase come fallback. Per le didascalie generate, trigger_phrase entra nel prompt di captioning e viene normalizzato nei sidecar risultanti secondo la convenzione di captioning LoRA selezionata. |
| `steps` |  | integer | no | `100` | 50 – 10,000 | Numero di passaggi per l'addestramento LoRA. |
| `learning_rate` |  | number | no | `0.0005` | 0.000001 – 0.01 | Learning rate del fattore B di LoRA, applicato anche ad A se `lora_a_lr_multiplier=1`; con muon, rappresenta il rate equivalente ad AdamW scalato da match_rms_adamw. Per una pianificazione costante, Krea consiglia l'intervallo 3e-4–7e-4. |
| `resolution` |  | enum (integer) | no | `768` | `768`, `1024` | Risoluzione di training: 1024 privilegia il massimo dettaglio fine, mentre 768 offre un training circa 1.7x più veloce, un uso ridotto della VRAM e una qualità dello stile quasi identica. Le immagini vengono ridimensionate per coprire il quadrato di destinazione, quindi ritagliate al centro. |
| `debug_dataset` |  | boolean | no | `false` |  | Opzione di download del dataset che espone esattamente i dati usati per il training, impacchettati come archivio dopo il completamento delle didascalie mancanti o la generazione automatica delle didascalie. |

## Example input

Required parameters only:

```json
{
  "images_data_url": "https://content.doitong.com/images/showcase/v3/gen_fox_cartoon.jpg"
}
```

A fuller example:

```json
{
  "images_data_url": "https://content.doitong.com/images/showcase/v3/gen_fox_cartoon.jpg"
}
```

## Run it

Set `DOITONG_API_KEY` to a key from https://doitong.com/developers/dashboard#keys. `?wait=20` returns the finished request in the same call when it is ready within 20 seconds; otherwise poll `GET /requests/{id}` (it also accepts `?wait=`), or pass `webhook_url`.

### curl

```bash
# Run the model; wait up to 20 s for the finished result in the same call
curl -X POST "https://api.doitong.com/v1/run/krea-2-trainer?wait=20" \
  -H "Authorization: Bearer $DOITONG_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "images_data_url": "https://content.doitong.com/images/showcase/v3/gen_fox_cartoon.jpg"
  }'

# Still "queued" or "running"? Poll with the id from the response:
curl "https://api.doitong.com/v1/requests/REQUEST_ID?wait=20" \
  -H "Authorization: Bearer $DOITONG_API_KEY"
```

### JavaScript

```javascript
const API = "https://api.doitong.com/v1";
const headers = {
  "Authorization": "Bearer " + process.env.DOITONG_API_KEY,
  "Content-Type": "application/json",
};

// Run the model; wait up to 20 s for the finished result in the same call
let res = await fetch(API + "/run/krea-2-trainer?wait=20", {
  method: "POST",
  headers,
  body: JSON.stringify({
    "images_data_url": "https://content.doitong.com/images/showcase/v3/gen_fox_cartoon.jpg"
  }),
});
let request = await res.json();
if (!res.ok) throw new Error(request.error.code + ": " + request.error.message);

// Not finished yet: long-poll until it is
while (request.status === "queued" || request.status === "running") {
  res = await fetch(API + "/requests/" + request.id + "?wait=20", { headers });
  request = await res.json();
}
if (request.status !== "succeeded") throw new Error(JSON.stringify(request.error));
console.log(request.outputs.map((o) => o.url), request.cost.credits + " API credits ($" + request.cost.usd + ")");
```

### Python

```python
import os
import requests

API = "https://api.doitong.com/v1"
headers = {"Authorization": "Bearer " + os.environ["DOITONG_API_KEY"]}

payload = {
    "images_data_url": "https://content.doitong.com/images/showcase/v3/gen_fox_cartoon.jpg",
}

# Run the model; wait up to 20 s for the finished result in the same call
r = requests.post(API + "/run/krea-2-trainer", params={"wait": 20}, json=payload, headers=headers)
request = r.json()
if not r.ok:
    raise RuntimeError(request["error"]["code"] + ": " + request["error"]["message"])

# Not finished yet: long-poll until it is
while request["status"] in ("queued", "running"):
    request = requests.get(API + "/requests/" + request["id"], params={"wait": 20}, headers=headers).json()
if request["status"] != "succeeded":
    raise RuntimeError(request["error"])
print([o["url"] for o in request["outputs"]], request["cost"]["credits"], "API credits", "$" + request["cost"]["usd"])
```

### MCP

The MCP server at `https://api.doitong.com/v1/mcp` exposes the tools `search_models`, `get_model`, `quote`, `run`, `get_request`, `cancel_request`, `list_requests`, `create_upload` and `get_account`. A `run` call for this model:

```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "run",
    "arguments": {
      "model_id": "krea-2-trainer",
      "input": {
        "images_data_url": "https://content.doitong.com/images/showcase/v3/gen_fox_cartoon.jpg"
      },
      "wait_seconds": 20
    }
  }
}
```

## Response

Every call returns a request object. `status` is one of `queued`, `running`, `succeeded`, `failed`, `canceled`. `outputs` lists the generated files with URLs on our storage (kept forever, or 7 days with `?storage=temporary`; `storage` says which); `output` is the model's own result object with the same URLs.

```json
{
  "id": "req_8k2m4q7x1z9c3v5b6n0p2r4t",
  "object": "request",
  "model": "krea-2-trainer",
  "status": "succeeded",
  "created_at": "2026-10-20T12:00:00Z",
  "started_at": "2026-10-20T12:00:02Z",
  "completed_at": "2026-10-20T12:00:09Z",
  "queue_position": null,
  "cost": {
    "credits": 300,
    "usd": "0.30",
    "charged": 300,
    "refunded": 0
  },
  "input": {
    "images_data_url": "https://content.doitong.com/images/showcase/v3/gen_fox_cartoon.jpg"
  },
  "output": {
    "lora_file": {
      "url": "https://content.doitong.com/dev-api/2026/10/20/req_8k2m4q7x1z9c3v5b6n0p2r4t/0-output.jpeg",
      "content_type": "image/jpeg",
      "file_name": "0-output.jpeg",
      "file_size": 1843200
    },
    "config_file": {
      "url": "https://content.doitong.com/dev-api/2026/10/20/req_8k2m4q7x1z9c3v5b6n0p2r4t/0-output.jpeg",
      "content_type": "image/jpeg",
      "file_name": "0-output.jpeg",
      "file_size": 1843200
    },
    "prompt_enhancement_system_prompt": "string",
    "debug_dataset": {
      "url": "https://content.doitong.com/dev-api/2026/10/20/req_8k2m4q7x1z9c3v5b6n0p2r4t/0-output.jpeg",
      "content_type": "image/jpeg",
      "file_name": "0-output.jpeg",
      "file_size": 1843200
    }
  },
  "outputs": [
    {
      "type": "file",
      "url": "https://content.doitong.com/dev-api/2026/10/20/req_8k2m4q7x1z9c3v5b6n0p2r4t/0-output.jpeg",
      "content_type": "image/jpeg",
      "file_size": 1843200,
      "stored": true
    }
  ],
  "error": null,
  "metrics": {
    "inference_seconds": 6.8
  },
  "urls": {
    "self": "https://api.doitong.com/v1/requests/req_8k2m4q7x1z9c3v5b6n0p2r4t",
    "cancel": "https://api.doitong.com/v1/requests/req_8k2m4q7x1z9c3v5b6n0p2r4t/cancel"
  }
}
```

### Output fields (`output`)

| Field | Type | Description |
|---|---|---|
| `lora_file` | object {url, content_type, file_name, file_size} | URL di download dei pesi LoRA Krea-2 addestrati in formato safetensors. |
| `config_file` | object {url, content_type, file_name, file_size} | Link a un file JSON contenente la configurazione di training/LoRA. |
| `prompt_enhancement_system_prompt` | string \| null | Prompt di sistema restituito con `auto_captioning` abilitato, utile per scrivere o perfezionare prompt di inferenza usando la stessa convenzione di captioning dei dati di training con didascalie automatiche. |
| `debug_dataset` | object {url, content_type, file_name, file_size} \| null | Download dei dati di training fornito quando `debug_dataset` è abilitato: un archivio contenente il dataset preparato e le relative didascalie generate automaticamente. |

## Related models

- [Krea 2 Text to Image Turbo LoRA](https://doitong.com/developers/models/krea-2/turbo/lora.md) (`krea-2/turbo/lora`): $0.01 / megapixel
- [Krea 2 Turbo](https://doitong.com/developers/models/krea-2/turbo.md) (`krea-2/turbo`): $0.008 / megapixel
- [ERNIE-Image Trainer](https://doitong.com/developers/models/ernie-image-trainer.md) (`ernie-image-trainer`): $0.0012 / step
- [FLUX 2 \[klein\] 9b Base Trainer](https://doitong.com/developers/models/flux-2-klein-9b-base-trainer.md) (`flux-2-klein-9b-base-trainer`): $0.0043 / step
- [Flux 2 Klein 9B Base Trainer](https://doitong.com/developers/models/flux-2-klein-9b-base-trainer/edit.md) (`flux-2-klein-9b-base-trainer/edit`): $0.0139 / step
- [FLUX 2 Trainer](https://doitong.com/developers/models/flux-2-trainer.md) (`flux-2-trainer`): $0.0064 / step
- [FLUX 2 Trainer Edit](https://doitong.com/developers/models/flux-2-trainer/edit.md) (`flux-2-trainer/edit`): $0.0389 / step
- [FLUX 2 Trainer V2](https://doitong.com/developers/models/flux-2-trainer-v2.md) (`flux-2-trainer-v2`): $0.0064 / step

## Links

- HTML page: https://doitong.com/developers/models/krea-2-trainer
- All models and guides for AI agents: https://doitong.com/developers/llms.txt
- OpenAPI 3.1 with this model's typed request body: https://doitong.com/developers/openapi.json?model=krea-2-trainer
- Errors: https://doitong.com/developers/docs/errors
- Model details as JSON: https://api.doitong.com/v1/models/krea-2-trainer