Prevent duplicate Vivideo video creations (and duplicate charges) when retrying or when multiple steps might submit the same request. Use whenever creating a video in a retryable or automated flow.
Verify a Vivideo account can afford a generation before creating it — check premium status and credit balance, estimate cost, and route the user to upgrade or buy credits if needed. Use before any video creation.
Create a Vivideo video by calling a specific model directly (manual mode) — Veo, Sora, Kling, Seedance, etc. Use when the user names a model or wants image-to-video / precise control over resolution, duration, aspect ratio.
Generate a Vivideo AI video from a text prompt (auto mode) end-to-end — check credits, create, wait safely, and return the finished MP4. Use when a user asks to make/generate a video from a description.
Correctly handle Vivideo API errors — rate limits, insufficient credits, premium requirements, duplicates, and retries. Use whenever a Vivideo tool/command returns an error.
Diagnose and recover from a failed or stuck Vivideo video generation. Use when a video's status is failed, or a render seems stuck, or a create call returned an error.
Wait for a Vivideo render to finish without over-polling or blocking forever, and decide between polling and webhooks. Use after creating a video, or whenever tracking a video to completion.
AI video generation via the Vivideo API: auto or manual mode, avatars, voices, brand kits. Runs locally from the @vivideo/mcp npm package. Needs 1 environment variable to run.