Hailuo Video Generator · MiniMax 2.3

Create a short Hailuo 2.3 video from text or a starting image, with an optional image caption. Choose 6 or 10 seconds at 768P, or 6 seconds at 1080P. After successful generation, download and save the video from its returned link.

Your AI

Use with your AI

For tasks with several steps, continue in the AI you already use.

CodexClaude CodeCursorWindsurfCline

View instructions

Use Toolmu Tool minimax-video-hailuo-2-3-minimax, ID 7e012180-7334-4a3a-ac05-b015f3cfa4ea. Reuse the existing Toolmu connection; when needed, follow https://toolmu.com/SKILL.md. Read this exact Tool’s current availability, inputs and customer credit ceiling, and honor the same account’s established authorization, balance and budget. Use MiniMax-Hailuo-2.3. For text-only generation, obtain a nonempty prompt of 1–2,000 characters; for image-to-video, obtain a compatible public image URL or Base64 Data URL, with an optional caption of up to 2,000 characters. Check the image format, size, short edge and aspect ratio. Choose 768P for 6 or 10 seconds, or 1080P for 6 seconds; omitted settings default to 768P and 6 seconds. Prompt optimization defaults on; fast_pretreatment defaults off and only affects optimization when it is enabled. Use supported camera commands such as [Pan left] when requested. Do not add ending-frame, aspect-ratio or reference-video/audio controls. Follow Toolmu’s authorized run and spending workflow and recover the original request if submission is uncertain. After success, use the returned download link to promptly save the video. Keep file and generation-task identifiers as returned, distinguish reported byte metadata from an actual download size, and report the actual status and confirmed charge. Task acceptance alone is not a finished video.

Task:
Make a short video of [a wooden toy train crossing a grassy meadow in soft morning light]. Have the camera [slowly move from right to left]. Choose [6 seconds at standard picture quality].

Paste into your connected AI's conversation.

Connection guide

Up to 0.672 credits / call

The final charge is based on confirmed usage. Unused reserved credits return to your available balance.

MiniMax

Hailuo Video Generator · MiniMax 2.3

You receive

Provide a scene prompt or a starting image to generate a short MiniMax-Hailuo-2.3 video. Select 6 or 10 seconds at 768P, or 6 seconds at 1080P. A successful result supplies a video download link, file and generation-task identifiers, filename, pixel dimensions and reported file-size metadata. Text can describe the action and camera movement; a starting image can be used with or without a caption.

Start with your task

Make a short video of [a wooden toy train crossing a grassy meadow in soft morning light]. Have the camera [slowly move from right to left]. Choose [6 seconds at standard picture quality].

What to know

A 10-second clip requires 768P; 1080P supports 6 seconds only. The image input sets the starting frame and provides no ending-frame control. A submitted task must finish successfully before a video is available. Save the finished file promptly from its download address.

Inputs & outputs

Inspect the current inputs, availability, and maximum charge before starting.

Input schema

Input schema
Field / Path	Type	Requirement	Meaning
Video-generation inputs`body`	`object`	Optional	Request settings for text-to-video or one starting image. The chosen input branch determines whether a nonempty prompt or image is needed.
Generation model`body.model`allOf · 1	`string`	—	Fixed to MiniMax-Hailuo-2.3 and used by default if omitted. This Tool does not select another Hailuo model through this field. Default `MiniMax-Hailuo-2.3` Fixed value `MiniMax-Hailuo-2.3`
Automatic prompt refinement`body.prompt_optimizer`allOf · 1	`boolean`	—	Boolean, default true. Allows automatic optimization of the written prompt. Set false when you want closer control over its supplied wording. Default `true`
Faster prompt preprocessing`body.fast_pretreatment`allOf · 1	`boolean`	—	Boolean, default false. Reduces optimization-stage time when prompt_optimizer is true; it has no optimization effect when that option is off. It does not select the Hailuo Fast model. Default `false`
Scene and camera description`body.prompt`allOf · 2 · anyOf · 1	`string`	—	For text alone, provide 1–2,000 characters describing the scene and motion. With a valid starting image, the caption is optional and may be omitted or empty, up to 2,000 characters. Supports camera commands such as [Pan left] and [Push in]. Minimum length `1` Maximum length `2000`
Scene and camera description`body.prompt`allOf · 2 · anyOf · 2	`string`	—	For text alone, provide 1–2,000 characters describing the scene and motion. With a valid starting image, the caption is optional and may be omitted or empty, up to 2,000 characters. Supports camera commands such as [Pan left] and [Push in]. Maximum length `2000`
Starting-frame image`body.first_frame_image`allOf · 2 · anyOf · 2	`string`	—	The image that begins the generated scene; required for image-to-video and not needed for text alone. Use a public URL or Base64 Data URL. JPG/PNG/WebP, less than 20 MB, short edge greater than 300 pixels and aspect ratio between 2:5 and 5:2. Its caption is optional. Minimum length `1`
Requested video length`body.duration`allOf · 3 · anyOf · 1	`number`	—	Duration in seconds, default 6. Choose 6 or 10 at 768P; 1080P accepts only 6. These are discrete choices, not a continuous range. Default `6`
Generated-video resolution`body.resolution`allOf · 3 · anyOf · 1	`string`	—	768P by default, or 1080P. A 6-second video can use either; a 10-second video must use 768P. Default `768P` Fixed value `768P`
Requested video length`body.duration`allOf · 3 · anyOf · 2	`number`	—	Duration in seconds, default 6. Choose 6 or 10 at 768P; 1080P accepts only 6. These are discrete choices, not a continuous range. Default `6` Fixed value `6`
Generated-video resolution`body.resolution`allOf · 3 · anyOf · 2	`string`	—	768P by default, or 1080P. A 6-second video can use either; a 10-second video must use 768P. Fixed value `1080P`

View raw JSON

{
  "type": "object",
  "required": [],
  "properties": {
    "body": {
      "allOf": [
        {
          "type": "object",
          "properties": {
            "model": {
              "type": "string",
              "const": "MiniMax-Hailuo-2.3",
              "default": "MiniMax-Hailuo-2.3",
              "description": "Fixed to MiniMax-Hailuo-2.3."
            },
            "prompt_optimizer": {
              "type": "boolean",
              "default": true,
              "description": "Auto-optimize the prompt. Default true."
            },
            "fast_pretreatment": {
              "type": "boolean",
              "default": false,
              "description": "Reduce optimization time when prompt_optimizer is on."
            }
          }
        },
        {
          "anyOf": [
            {
              "type": "object",
              "required": [
                "prompt"
              ],
              "properties": {
                "prompt": {
                  "type": "string",
                  "maxLength": 2000,
                  "minLength": 1,
                  "description": "Text description of the video (max 2000). Supports [camera command] syntax, e.g. [Push in], [Pan left], [Zoom out]."
                }
              }
            },
            {
              "type": "object",
              "required": [
                "first_frame_image"
              ],
              "properties": {
                "prompt": {
                  "type": "string",
                  "maxLength": 2000,
                  "description": "Optional caption for image-to-video. Supports [camera command] syntax."
                },
                "first_frame_image": {
                  "type": "string",
                  "minLength": 1,
                  "description": "Starting-frame image (image-to-video). Public URL or base64 data URL (data:image/jpeg;base64,...). JPG/PNG/WebP, <20MB, short edge >300px, aspect ratio between 2:5 and 5:2."
                }
              }
            }
          ]
        },
        {
          "anyOf": [
            {
              "type": "object",
              "properties": {
                "duration": {
                  "anyOf": [
                    {
                      "type": "number",
                      "const": 6
                    },
                    {
                      "type": "number",
                      "const": 10
                    }
                  ],
                  "default": 6,
                  "description": "6 or 10 seconds (768P)."
                },
                "resolution": {
                  "type": "string",
                  "const": "768P",
                  "default": "768P"
                }
              }
            },
            {
              "type": "object",
              "required": [
                "resolution"
              ],
              "properties": {
                "duration": {
                  "type": "number",
                  "const": 6,
                  "default": 6,
                  "description": "1080P supports 6s only."
                },
                "resolution": {
                  "type": "string",
                  "const": "1080P"
                }
              }
            }
          ]
        }
      ],
      "$schema": "https://json-schema.org/draft/2020-12/schema",
      "~standard": {
        "vendor": "zod",
        "version": 1,
        "jsonSchema": {}
      }
    }
  },
  "additionalProperties": false
}

Result schema

Output fields have not been documented yet.

Additional notes

Text-to-video starts from your description of a scene, action and camera movement. Image-to-video instead uses one image as the opening frame; an optional caption can describe how the scene should move. A valid image request can omit the caption or leave it empty. The starting image guides the opening scene without adding an ending-frame control.

Camera commands can be placed in the prompt: [Truck left], [Truck right], [Pan left], [Pan right], [Push in], [Pull out], [Pedestal up], [Pedestal down], [Tilt up], [Tilt down], [Zoom in], [Zoom out], [Shake], [Tracking shot], [Static shot]. Put several commands in one bracket group for simultaneous movement; MiniMax recommends no more than three together. Separate groups can describe movements in sequence, and ordinary descriptions of camera motion also work.

Prompt optimization is on by default and automatically refines the written prompt. Turn it off when you want closer control over the wording supplied for generation. Fast pretreatment defaults off; when enabled alongside prompt optimization, it reduces the optimization stage’s time. It keeps the same Hailuo 2.3 model and does not define the overall video-completion time.

A successful result provides download_url, the address for retrieving the generated video, and filename, its suggested file name. file_id identifies that video file; task_id identifies the supplier generation task, separate from the Toolmu run. Keep both identifiers as strings. video_width and video_height describe the video dimensions in pixels. bytes is the supplier’s reported file-size metadata in bytes; a reported zero alone does not prove an empty video or establish its downloaded size. Save the video promptly from the returned link.

MiniMax Hailuo text generation and camera controls

MiniMax Hailuo image input and optional caption

MiniMax generated-video file retrieval

Questions about this tool.

Inputs, results and details to know before running it.

Get connected

A valid starting-image request can omit its caption or leave it empty. Text-only generation needs a nonempty 1–2,000-character prompt. Use a caption when you want to describe motion in the image.

Hailuo Video Generator · MiniMax 2.3

Hailuo Video Generator · MiniMax 2.3

You receive

Start with your task

What to know

Inputs & outputs

Input schema

Result schema

Describe a scene or add a starting picture

Choose a video length and picture clarity

Save the finished video

Questions about this tool.

You receive

Start with your task

What to know

Inputs & outputs

Input schema

Result schema

Describe a scene or add a starting picture

Choose a video length and picture clarity

Save the finished video

Questions about this tool.

Does an image request need a written prompt?

Can I generate 10 seconds at 1080P?

Can I control the ending frame?

Does fast pretreatment switch to a Fast model?