# Flux 3 Action API

> 机器人策略模型，根据相机帧、关节状态和指令预测 SO-101 机械臂动作块

该模型不返回媒体，而是返回包含预测 `actions`、控制频率、列标签以及在请求新动作块前要执行的前导步骤数（`execute_steps`）的 JSON。必填输入包括文本 `prompt`（最多 2000 个字符）、`scene_image_url`、`wrist_image_url` 和包含 6 个值的 `state`。可选的历史输入（`state_history`、`command_history`、历史图像）用于提供时间上下文。`num_inference_steps` 范围为 1 到 8，`guidance_scale` 范围为 1 到 10。

- Model id: `flux-3-action/so101`
- Made by: Black Forest Labs
- Category: Video to Video
- Output: json
- Estimated price per run: $0.003 / run (cheapest run $0.003 (3 API credits))
- Run: `POST https://api.doitong.com/v1/run/flux-3-action/so101`
- Quote: `POST https://api.doitong.com/v1/quote/flux-3-action/so101`
- Docs page: https://doitong.com/zh/developers/models/flux-3-action/so101

## Pricing

Price: **$0.003 / run**. Estimated from the typical cost of one run. The quote is exactly what the run charges.

The full example input below costs $0.003 = 3 API credits (1 request × 3 API credits).

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/flux-3-action/so101` 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 |
|---|---|---|---|---|---|---|
| `prompt` |  | string | yes |  | 1 – 2000 characters | 以自然语言表达的机器人任务指令。 |
| `scene_image_url` |  | string (image URL) | yes |  |  | 左侧图块中的图像，显示固定场景摄像头在调整为 256x256 后的当前帧。 |
| `wrist_image_url` |  | string (image URL) | yes |  |  | 当前腕部摄像头帧，位于右侧图块中，并调整为 256x256。 |
| `state` |  | array of number | yes |  | 6 – 6 items | 按顺序排列的测量状态值：肩部旋转和抬升、肘部弯曲、腕部弯曲和旋转，单位均为度；随后是夹爪位置百分比。 |
| `state_history` |  | array of array of number \| null | no |  | 7 – 7 items | 测量状态历史记录；新 episode 中省略，否则需要 `command_history`。按时间顺序提供 `state` 前紧邻的 7 个控制 tick 的测量值，频率为 30 Hz。 |
| `command_history` |  | array of array of number \| null | no |  | 8 – 8 items | 需要 `state_history` 的绝对指令历史记录，最后一行是最近发送的指令。各行按 8 个观测 tick 的时间顺序排列，并指定每个 tick 生效的指令。 |
| `history_scene_image_url` |  | string (image URL) \| null | no |  |  | 需要 `state_history` 的历史场景帧输入；省略时使用当前帧。提供的帧代表 7 个 tick 之前最早的历史观测。 |
| `history_wrist_image_url` |  | string (image URL) \| null | no |  |  | 最早历史 tick 的腕部摄像头帧；需要 `state_history`。 |
| `num_inference_steps` |  | integer | no | `4` | 1 – 8 | Euler 去噪步数；检查点验证使用 4 步。 |
| `guidance_scale` |  | number | no | `3` | 1 – 10 | 视频和动作的无分类器引导强度；设为 1.0 可禁用 CFG。 |
| `seed` |  | integer \| null | no |  |  | 采样种子；DROID 的参考评估设置为 0，SO-101 为 42。未指定时随机选择种子。 |

File parameters (`scene_image_url`, `wrist_image_url`, `history_scene_image_url`, `history_wrist_image_url`) take a public `https://` URL, a `data:` URI up to 10 MB, or the `file_url` of an upload from `POST /files`.

## Example input

Required parameters only:

```json
{
  "prompt": "A red fox trots through fresh snow at sunrise, soft golden light, shallow depth of field",
  "scene_image_url": "https://content.doitong.com/images/showcase/v3/gen_fox_cartoon.jpg",
  "wrist_image_url": "https://content.doitong.com/images/showcase/v3/gen_fox_cartoon.jpg",
  "state": [
    -18.94,
    134.52,
    120.15,
    28.3,
    -32.17,
    27.02
  ]
}
```

A fuller example:

```json
{
  "prompt": "A red fox trots through fresh snow at sunrise, soft golden light, shallow depth of field",
  "scene_image_url": "https://content.doitong.com/images/showcase/v3/gen_fox_cartoon.jpg",
  "wrist_image_url": "https://content.doitong.com/images/showcase/v3/gen_fox_cartoon.jpg",
  "state": [
    -18.94,
    134.52,
    120.15,
    28.3,
    -32.17,
    27.02
  ]
}
```

## 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/flux-3-action/so101?wait=20" \
  -H "Authorization: Bearer $DOITONG_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "A red fox trots through fresh snow at sunrise, soft golden light, shallow depth of field",
    "scene_image_url": "https://content.doitong.com/images/showcase/v3/gen_fox_cartoon.jpg",
    "wrist_image_url": "https://content.doitong.com/images/showcase/v3/gen_fox_cartoon.jpg",
    "state": [
      -18.94,
      134.52,
      120.15,
      28.3,
      -32.17,
      27.02
    ]
  }'

# 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/flux-3-action/so101?wait=20", {
  method: "POST",
  headers,
  body: JSON.stringify({
    "prompt": "A red fox trots through fresh snow at sunrise, soft golden light, shallow depth of field",
    "scene_image_url": "https://content.doitong.com/images/showcase/v3/gen_fox_cartoon.jpg",
    "wrist_image_url": "https://content.doitong.com/images/showcase/v3/gen_fox_cartoon.jpg",
    "state": [
      -18.94,
      134.52,
      120.15,
      28.3,
      -32.17,
      27.02
    ]
  }),
});
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 = {
    "prompt": "A red fox trots through fresh snow at sunrise, soft golden light, shallow depth of field",
    "scene_image_url": "https://content.doitong.com/images/showcase/v3/gen_fox_cartoon.jpg",
    "wrist_image_url": "https://content.doitong.com/images/showcase/v3/gen_fox_cartoon.jpg",
    "state": [
        -18.94,
        134.52,
        120.15,
        28.3,
        -32.17,
        27.02,
    ],
}

# Run the model; wait up to 20 s for the finished result in the same call
r = requests.post(API + "/run/flux-3-action/so101", 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": "flux-3-action/so101",
      "input": {
        "prompt": "A red fox trots through fresh snow at sunrise, soft golden light, shallow depth of field",
        "scene_image_url": "https://content.doitong.com/images/showcase/v3/gen_fox_cartoon.jpg",
        "wrist_image_url": "https://content.doitong.com/images/showcase/v3/gen_fox_cartoon.jpg",
        "state": [
          -18.94,
          134.52,
          120.15,
          28.3,
          -32.17,
          27.02
        ]
      },
      "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": "flux-3-action/so101",
  "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": 3,
    "usd": "0.003",
    "charged": 3,
    "refunded": 0
  },
  "input": {
    "prompt": "A red fox trots through fresh snow at sunrise, soft golden light, shallow depth of field",
    "scene_image_url": "https://content.doitong.com/images/showcase/v3/gen_fox_cartoon.jpg",
    "wrist_image_url": "https://content.doitong.com/images/showcase/v3/gen_fox_cartoon.jpg",
    "state": [
      -18.94,
      134.52,
      120.15,
      28.3,
      -32.17,
      27.02
    ]
  },
  "output": {
    "actions": [
      [
        1.0
      ]
    ],
    "action_labels": [
      "string"
    ],
    "action_hz": 1.0,
    "execute_steps": 1,
    "seed": 42,
    "timings": {}
  },
  "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 |
|---|---|---|
| `actions` | array of array of number | 预测的动作数组，各行按时间顺序表示控制步骤；数值未裁剪，因此必须在机器人上执行关节、速度和工作空间限制。30 Hz 时，SO-101 提供 42x6 个绝对指令，单位为 deg 和夹爪 %；15 Hz 时，DROID 提供 32x8 个绝对关节目标，单位为 rad，另加夹爪闭合比例。 |
| `action_labels` | array of string | 对应动作列的名称。 |
| `action_hz` | number | 动作行的控制频率。 |
| `execute_steps` | integer | 在请求下一批动作前要执行的初始动作数量。 |
| `seed` | integer | 采样期间使用的随机种子。 |
| `timings` | object | 服务器的计时测量值，单位为秒。 |

## Related models

- [Flux 3 Draft Enhance](https://doitong.com/developers/models/blackforestlabs/flux-3/draft-enhance.md) (`blackforestlabs/flux-3/draft-enhance`): $3.48 / run
- [Flux 3 Extend Video](https://doitong.com/developers/models/blackforestlabs/flux-3/extend-video.md) (`blackforestlabs/flux-3/extend-video`): from $0.41 / second
- [Flux 3 Extend Video Draft](https://doitong.com/developers/models/blackforestlabs/flux-3/extend-video/draft.md) (`blackforestlabs/flux-3/extend-video/draft`): $0.12 / second
- [Flux 3 FAST Edit Video](https://doitong.com/developers/models/blackforestlabs/flux-3/edit-video.md) (`blackforestlabs/flux-3/edit-video`): $0.03 / second of input video
- [AMT Interpolation](https://doitong.com/developers/models/amt-interpolation.md) (`amt-interpolation`): $0.000575 / second of processing time
- [AnimateDiff Turbo · Video to Video](https://doitong.com/developers/models/fast-animatediff/turbo/video-to-video.md) (`fast-animatediff/turbo/video-to-video`): $0.00125 / second of processing time
- [AnimateDiff · Video to Video](https://doitong.com/developers/models/fast-animatediff/video-to-video.md) (`fast-animatediff/video-to-video`): $0.00125 / second of processing time
- [Auto-Captioner](https://doitong.com/developers/models/auto-caption.md) (`auto-caption`): $0.10 / video

## Links

- HTML page: https://doitong.com/developers/models/flux-3-action/so101
- 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=flux-3-action/so101
- Errors: https://doitong.com/developers/docs/errors
- Model details as JSON: https://api.doitong.com/v1/models/flux-3-action/so101