Skip to content
AutoVio Docs

MCP Workflows

MCP Workflows

Section titled “MCP Workflows”

This page shows example workflows using the AutoVio MCP tools to drive the video pipeline.

All examples assume:

  • The AutoVio MCP server is running and configured (see MCP Setup).
  • The MCP client can call tools with a JSON args object and display JSON text results.

Workflow 1 – Create project and work

Section titled “Workflow 1 – Create project and work”

1. Create a project

Section titled “1. Create a project”

Use autovio_projects_create with a project type for optimized prompts:

{
  "name": "MCP Demo Project",
  "projectType": "ecommerce",
  "knowledge": "We sell eco-friendly products for outdoor enthusiasts."
}

Project types: blank, saas, news, social, ecommerce, educational. Each comes with optimized prompts and style guide.

Copy the returned id as projectId.

2. Create a work

Section titled “2. Create a work”

Call autovio_works_create:

{
  "projectId": "proj_123",
  "name": "MCP Demo Video",
  "mode": "content_remix",
  "productName": "EcoBottle",
  "productDescription": "Eco-friendly reusable bottle",
  "sceneCount": 5,
  "videoDuration": 30
}

Copy the returned id as workId.

Workflow 2 – Generate scenario and scenes for a work

Section titled “Workflow 2 – Generate scenario and scenes for a work”

1. Generate scenario for the work

Section titled “1. Generate scenario for the work”

Use autovio_works_generate_scenario:

{
  "projectId": "proj_123",
  "workId": "work_123"
}

The response contains scenes for the work. The work in the backend is also updated.

2. Generate image and video for each scene

Section titled “2. Generate image and video for each scene”

For each scene index (0-based), call autovio_works_generate_scene:

{
  "projectId": "proj_123",
  "workId": "work_123",
  "sceneIndex": 0
}

Repeat for sceneIndex 1, 2, etc. Each call:

  • Generates an image and video for the scene.
  • Saves media under the work’s media endpoints.
  • Updates generatedScenes for that scene.

3. Inspect the work

Section titled “3. Inspect the work”

Optionally call autovio_works_get:

{
  "projectId": "proj_123",
  "workId": "work_123"
}

This returns the full work object, including scenes, generated media URLs, and editor state (if previously saved by the UI or API).

Workflow 3 – Use direct AI tools

Section titled “Workflow 3 – Use direct AI tools”

You can also use AI tools that are not tied to a specific work.

1. Analyze a reference video

Section titled “1. Analyze a reference video”

Encode a small video file as base64 and call autovio_ai_analyze_video:

{
  "videoBase64": "BASE64_CONTENT_HERE",
  "mode": "style_transfer",
  "analyzerPrompt": "Focus on camera movement and text overlays."
}

The result is an analysis object you can feed into scenario generation.

2. Generate a scenario

Section titled “2. Generate a scenario”

Use autovio_ai_generate_scenario:

{
  "intent": {
    "mode": "content_remix",
    "product_name": "EcoBottle",
    "product_description": "Eco-friendly reusable bottle",
    "target_audience": "Eco-conscious users",
    "language": "en",
    "video_duration": 30,
    "scene_count": 3
  },
  "analysis": { "...": "output from analyze_video" },
  "knowledge": "Brand focuses on sustainability.",
  "styleGuide": {
    "tone": "professional"
  }
}

The response contains a scenes array with prompts you can then use with autovio_ai_generate_image and autovio_ai_generate_video.

Workflow 4 – Apply a template to a work

Section titled “Workflow 4 – Apply a template to a work”

1. List templates

Section titled “1. List templates”

Call autovio_templates_list:

{
  "projectId": "proj_123"
}

Pick a templateId from the result.

2. Apply template to the work

Section titled “2. Apply template to the work”

Use autovio_works_apply_template:

{
  "projectId": "proj_123",
  "workId": "work_123",
  "templateId": "tmpl_123",
  "placeholderValues": {
    "product_name": "EcoBottle",
    "brand": "Acme Co."
  }
}

The response is the updated work with merged editor state (timeline, overlays, export settings).

Workflow 5 – Asset-based video generation

Section titled “Workflow 5 – Asset-based video generation”

Use project assets directly in video generation instead of AI-generated images.

1. List project assets

Section titled “1. List project assets”

Call autovio_assets_list:

{
  "projectId": "proj_123",
  "type": "image"
}

Note the id values for assets you want to use.

2. (Optional) Analyze assets for reference mode

Section titled “2. (Optional) Analyze assets for reference mode”

For reference mode, analyze assets to generate descriptions:

{
  "projectId": "proj_123",
  "assetIds": ["asset_1", "asset_2", "asset_3"]
}

Use autovio_assets_analyze_batch. Descriptions help AI generate similar-looking images.

3. Create work with assets

Section titled “3. Create work with assets”

Use autovio_works_create with asset selection:

{
  "projectId": "proj_123",
  "name": "Product Showcase",
  "mode": "content_remix",
  "productName": "EcoBottle",
  "selectedAssetIds": ["asset_1", "asset_2", "asset_3"],
  "assetUsageMode": "direct"
}

Asset usage modes:

  • direct — Use actual asset images, skip AI image generation. Scene count auto-adjusts to asset count.
  • reference — AI generates new images matching asset style (requires analyzed assets).

4. Generate scenario and scenes

Section titled “4. Generate scenario and scenes”

With direct mode, scenario generation creates scenes matching your asset count. Image generation is skipped; only video (animation) is generated from your actual photos.

See also

Section titled “See also”