VideoOrchestrator

The main orchestration class for managing video generation workflows.

Constructor

VideoOrchestrator(provider: VideoProvider)

Parameters

provider

VideoProvider - The video generation provider to use

Example

from ministudio import VideoOrchestrator, VertexAIProvider

provider = VertexAIProvider()
orchestrator = VideoOrchestrator(provider)

Methods

generate_shot()

Generate a single video shot.

async def generate_shot(
    request: VideoGenerationRequest
) -> VideoGenerationResult

Parameters

request

VideoGenerationRequest - The generation request configuration

Returns

VideoGenerationResult

The result containing the generated video

Example

result = await orchestrator.generate_shot(
    VideoGenerationRequest(
        prompt="A scientist in a laboratory",
        duration_seconds=8
    )
)

if result.success:
    print(f"Video generated: {result.video_path}")
else:
    print(f"Error: {result.error}")

generate_shot() is an async function. Use it with await or asyncio.run().

generate_sequence()

Generate a multi-shot sequence with automatic continuity.

async def generate_sequence(
    shots: List[VideoGenerationRequest]
) -> List[VideoGenerationResult]

Parameters

shots

List[VideoGenerationRequest] - List of shot requests to generate

Returns

List[VideoGenerationResult]

Results for each shot in the sequence

Example

shots = [
    VideoGenerationRequest(
        prompt="A researcher enters the laboratory",
        duration_seconds=5
    ),
    VideoGenerationRequest(
        prompt="The researcher examines equipment",
        duration_seconds=6
    ),
    VideoGenerationRequest(
        prompt="The researcher makes notes",
        duration_seconds=4
    )
]

results = await orchestrator.generate_sequence(shots)
video_paths = [r.video_path for r in results if r.success]

The orchestrator automatically handles frame extraction and state management for continuity between shots.

get_state()

Get the current state of the orchestrator.

def get_state() -> OrchestratorState

Returns

OrchestratorState

Current state including shot history and continuity data

reset()

Reset the orchestrator state for a new sequence.

def reset() -> None

Example

# Finish first sequence
results1 = await orchestrator.generate_sequence(shots1)

# Reset for new sequence
orchestrator.reset()

# Start new sequence
results2 = await orchestrator.generate_sequence(shots2)

Methods

generate_shot()

Generate a single video shot.

async def generate_shot(
    request: VideoGenerationRequest
) -> VideoGenerationResult

Parameters

request

VideoGenerationRequest - The generation request configuration

Returns

VideoGenerationResult

The result containing the generated video

Example

result = await orchestrator.generate_shot(
    VideoGenerationRequest(
        prompt="A scientist in a laboratory",
        duration_seconds=8
    )
)

if result.success:
    print(f"Video generated: {result.video_path}")
else:
    print(f"Error: {result.error}")

generate_shot() is an async function. Use it with await or asyncio.run().

shots = [ VideoGenerationRequest( prompt="A researcher enters the laboratory", duration_seconds=5 ), VideoGenerationRequest( prompt="The researcher examines equipment", duration_seconds=6 ), VideoGenerationRequest( prompt="The researcher makes notes", duration_seconds=4 ) ] results = await orchestrator.generate_sequence(shots) video_paths = [r.video_path for r in results if r.success]

# Finish first sequence results1 = await orchestrator.generate_sequence(shots1) # Reset for new sequence orchestrator.reset() # Start new sequence results2 = await orchestrator.generate_sequence(shots2)