Connect MiniMax to Your AI Agent

Describe a scene and its output settings to create a video task, then check when it finishes.

13 tools API key

Try a MiniMax Task

Copy a prompt into your agent with the oo Skill installed. Add your page links and source notes before running it.

oo Skill MiniMax

Create and Check an H3 Video

Generate a short scene with chosen dimensions and check the task result.

/oo MiniMax: Create a MiniMax-H3 video showing [scene], 6 seconds long, at 768P with a 16:9 ratio. Query the task until it finishes or reports failure, and show the result.

Draft Text with a Selected Model

Generate text from supplied material with a chosen available model.

/oo MiniMax: Using MiniMax model [model ID], draft [text type] from [source material] for [audience].

Start Working with MiniMax

  1. Install OOMOL

    Set up oo CLI and the oo Skill in your agent.

    Install OOMOL →
  2. Connect MiniMax

    Connect a MiniMax API key for the global or China region, or use Marketplace for H3 creation and querying.

    Connect MiniMax →
  3. Give Your Agent a Task

    Use /oo with the pages to read, the changes to make, and where to save the result.

    Connection Guide →

Choose Your Agent

  • A MiniMax key matching its API region for direct access; H3 tasks need a prompt, resolution, and duration.
  • Video creation returns a task ID. Wait for a completed result before treating the video as ready.

MiniMax Tools

The connection exposes MiniMax text responses, speech synthesis, and asynchronous video generation. Marketplace access covers H3 video creation and querying; other operations use a MiniMax API key.

Create and Inspect Video Tasks

Create H3 or legacy video tasks and retrieve their status; H3 also supports listing and deletion.

Generate Text and Speech

Inspect models, generate text, estimate input tokens, and synthesize speech.

13 tools

List Models

List OpenAI-compatible MiniMax models available to the API key.

minimax.list_models

Retrieve Model

Retrieve OpenAI-compatible metadata for one MiniMax model.

minimax.retrieve_model

Create Response

Create a non-streaming MiniMax response using the OpenAI Responses API shape.

minimax.create_response

Estimate Input Tokens

Estimate MiniMax response input tokens without invoking the model.

minimax.estimate_input_tokens

Create Video Generation V2

Create a MiniMax H3 video generation task with text, image, video, or audio content.

minimax.create_video_generation_v2

Text To Video

Create a MiniMax asynchronous text-to-video generation task.

minimax.text_to_video

Image To Video

Create a MiniMax asynchronous image-to-video task from a public URL or data URI first frame.

minimax.image_to_video

Query Video Generation

Query a MiniMax video generation task and return its status and file identifier.

minimax.query_video_generation

Query Video Generation V2

Query a MiniMax H3 video generation task.

minimax.query_video_generation_v2

List Video Generation V2

List MiniMax H3 video generation tasks.

minimax.list_video_generation_v2

Delete Video Generation V2

Delete a MiniMax H3 video generation task.

minimax.delete_video_generation_v2

Download Video

Download a generated MiniMax video and store it in connector transit storage.

minimax.download_video

Text To Audio

Synthesize text into audio with the MiniMax T2A v2 API.

minimax.text_to_audio

Tools are available through the OOMOL marketplace. Usage charges may apply.

Frequently Asked Questions

What H3 output settings can I choose?

H3 accepts 4–15 seconds and 768P or 2K resolution, with supported aspect ratios including 16:9 and 9:16.

Can I provide reference media?

H3 accepts text with image, video, or audio references. Identify each reference’s purpose and provide a usable media URL.

How are legacy video files delivered?

Successful legacy task queries download the generated file into connector transit storage. This requires file transit to be configured.

Connect MiniMax

About MiniMax ↗