← 返回
未分类 中文

Freebeat Music Video Generator

Generate AI music videos from any MCP client. Turn text prompts into cinematic music videos with multiple styles and modes. Existing features include charact...
将文字提示转为多种风格与模式的电影级AI音乐视频,支持任何MCP客户端调用,现有功能包括角色...
d3ter28 d3ter28 来源
未分类 clawhub v1.0.2 1 版本 100000 Key: 无需
★ 0
Stars
📥 363
下载
💾 0
安装
1
版本
#latest

概述

Freebeat MCP Workflows

Use this skill for reliable Freebeat MCP task execution with correct tool order and async handling.

Quick Start

  1. Confirm the user goal:
    • generate_effect (template-based effect generation)
    • generate_music_video (MV generation from uploaded audio)
  2. Gather required inputs.
  3. Execute the correct workflow.
  4. Poll task status until terminal state.
  5. Fetch result only after completion.

Tool Map

  • upload_audio: Upload local audio or import from URL. Returns music_id.
  • upload_image: Upload one or two images. Returns image_urls.
  • list_effects: Discover templates and defaults for effect generation.
  • generate_effect: Start async effect generation task.
  • generate_music_video: Start async MV generation task.
  • get_task_status: Poll async task state.
  • get_task_result: Fetch output only when status is completed.

Workflow A: Generate Effect

Preferred order:

  1. Call list_effects.
  2. Select effect_id and collect defaults:
    • default_music_id
    • image_url (default image)
  3. Optional overrides:
    • If user provides custom music, call upload_audio and use returned music_id.
    • If user provides custom image, call upload_image with one image and use image_urls[0].
  4. Call generate_effect with:
    • effect_id (required)
    • music_id (required)
    • prompt (required, trimmed length 1..2000)
    • reference_image_urls (required, exactly one URL)
    • watermark (optional, default false)
  5. Poll with get_task_status until completed or failed.
  6. On completed, call get_task_result and return video_url and cover_url.

Workflow B: Generate Music Video

Preferred order:

  1. Call upload_audio first and capture music_id.
  2. Optional image references:
    • Call upload_image and pass returned image_urls to generate_music_video.reference_image_urls.
  3. Call generate_music_video with:
    • music_id (required)
    • prompt (required, trimmed length 1..2000)
    • Optional: mv_type, style, aspect_ratio, resolution, watermark, start_ms, end_ms, reference_image_urls
  4. Poll with get_task_status until terminal state.
  5. On completed, call get_task_result and return video_url and cover_url.

Async Handling Rules

  • Treat task submission as asynchronous.
  • Continue polling while status indicates in-progress work (for example pending).
  • Status values may be lowercase (pending, completed, failed) and some older systems may use uppercase variants.
  • Do not call get_task_result before status is completed.
  • If get_task_result returns TASK_NOT_COMPLETED, continue polling.
  • On failed, report error_message, error_code, and any backend message from status response.

Input Validation Checklist

Before generation calls, verify:

  • Prompt is present and trimmed length is within 1..2000.
  • For generate_effect, reference_image_urls contains exactly one image URL.
  • For generate_effect, music_id is either default_music_id from list_effects or from upload_audio.
  • For generate_music_video, music_id comes from upload_audio.
  • For upload_audio, provide exactly one of file_path or url.

Response Format Guidelines

When reporting progress or final outputs, include:

  • Task type (effect_generation or music_video_generation)
  • task_id
  • Current status
  • On completion: video_url and cover_url
  • On failure: failure code/message and next fix step

Defaults and Practical Guidance

  • Use list_effects defaults unless the user explicitly asks to override.
  • For effect generation, default image is list_effects.image_url.
  • For MV generation, default values are acceptable unless user specifies otherwise.
  • Keep polling cadence steady and avoid tight loops.

Example Decision Logic

If user asks "make an effect from template":

  • Use Workflow A.

If user asks "make a music video from this song":

  • Use Workflow B.

If user provides only a prompt without media:

  • For effect: use list_effects defaults.
  • For MV: request audio input (or URL) and run upload_audio first.

版本历史

共 1 个版本

  • v1.0.2 当前
    2026-05-07 11:53 安全 安全

安全检测

腾讯云安全 (Keen)

安全,无风险
查看报告

腾讯云安全 (Sanbu)

安全,无风险
查看报告

🔗 相关推荐

design-media

Video Frames

steipete
使用 ffmpeg 从视频中提取帧或短片。
★ 137 📥 53,441
design-media

Openai Whisper

steipete
使用 Whisper CLI 进行本地语音转文字(无需 API 密钥)
★ 337 📥 95,255
design-media

Nano Banana Pro

steipete
使用 Nano Banana Pro (Gemini 3 Pro Image) 生成或编辑图像。支持文生图、图生图及 1K/2K/4K 分辨率,适用于图像创建、修改及编辑请求,使用 --input-image 指定输入图像。
★ 439 📥 118,581